J. Cent. South Univ. Technol. (2009) 16: 0982-0986

DOI: 10.1007/s11771-009-0163-7

Non-invasive glucose measuring apparatus based on

conservation of energy method

CHEN Zhen-cheng(陈真诚), JIN Xing-liang(金星亮),

ZHU Jian-ming(朱健铭), WANG Di-ya(王弟亚), ZHANG Ting-ting(张婷婷)

(School of Info-physics and Geomatics Engineering, Central South University, Changsha 410083, China)

Abstract: A new non-invasive blood glucose measuring apparatus (NBGMA) made up of MSP430F149 SCM (single chip micyoco) was developed, which can measure blood glucose level (BGL) frequently, conveniently and painlessly. The hardware and software of this apparatus were designed, and detecting algorithms based on conservation of energy method (COEM) were presented. According to the law of conservation of energy that the energy derived by human body equals energy consumed by metabolism, and the relationship between convection, evaporation, radiation and the BGL was established. The sensor module was designed. 20 healthy volunteers were involved in the clinical experiment. The BGL measured by an automatic biochemical analyzer (ABA) was set as the reference. Regression analysis was performed to compare the conservation of energy method with the biochemical method, using the 20 data points with blood glucose concentrations ranging from 680 to 1 100 mg/L. Reproducibility was measured for healthy fasting volunteers. The results show that the means of BGL detected by NBGMA and ANA are very close to each other, and the difference of standard deviation (SD) is 24.7 mg/L. The correlative coefficient is 0.807. The coefficient of variation (CV) is 4% at 921.6 mg/L. The resultant regression is evaluated by the Clarke error grid analysis (EGA) and all data points are included in the clinically acceptable regions (region A: 100%, region B: 0%). Accordingly, it is feasible to measure BGL with COEM.

Key words: single chip; non-invasive measurement; blood glucose; conservation of energy method

1 Introduction

Diabetes mellitus, a chronic disease, usually results from inherent structural weaknesses of insulin or a diminished secretion of insulin into the bloodstream by Langerhans contained within the pancreas[1]. Precise and real time blood-glucose control is necessary and extremely important in the diabetes mellitus therapy. Current detecting technologies are invasive, which are time-consuming effort, painful, scarring, and not suitable for home measure[2-5]. Hence, it is desirable to develop a novel apparatus to satisfy the need of frequent detecting.

Non-invasive glucose measuring technologies have been extremely focused on for these years, which involve types such as: optical approaches and radiation approaches[6-8], fluid extraction approaches[9-11], metabolic heat conformation[12-13] and impedance spectroscopy[14]. The main study focuses on near-infrared (NIR) spectroscopy[15]. However, studies of blood glucose measurement using NIR spectroscopy have been disappointing. Meanwhile, due to the necessity of individual daily calibration and environmental interference, many hurdles remain before this product reaches the commercial marketplace. Then a novel non-invasive blood glucose measuring apparatus (NBGMA) based on conservation of energy method (COEM) was developed. This apparatus computes blood glucose level (BGL) by measuring parameters such as the thermal energy generated by metabolic reactions, oxygen saturation of hemoglobin and blood flow. Instead of traditional detecting method, this novel method makes it possible to detect BGL frequently, which is convenient and painless.

2.1 Principle

The process of substance metabolism in the entire human body is the process of energy metabolism[16]. When the oxygen supply is enough, the blood sugar, which is the main energy source substance, undergoes the reaction as follows:

C₆H₁₂O₆+O₂→H₂O+CO₂+ATP                  (1)

It can be seen that the energy produced by human body transforms into the heat generally except the work towards external. Physiology estimates that 80% of heat dissipates in different forms. The diversification of the concentration of the blood glucose leads to the change of metabolism in human body, and affects the variety of the physiological parameters, such as the radiation heat of the body surface. So CHO and KIM[16] set up the model under the following assumptions.

(1) The heat produced by human body and the heat consumed by human body are considered to be equal.

(2) When human is calm, the work towards external is zero.

(3) The heat produced human body can be described by the blood glucose concentration and the volume of oxygen supply.

(4) The volume of oxygen supply is determined by the blood hemoglobin concentration, the oxygen saturation of hemoglobin and the blood volume of capillary vessel.

(5) The main forms of heat dissipation of body are convection, radiation and evaporation.

According to the supposed theory stated above, the metabolic heat is the function of blood glucose concentration and oxygen supply. Oxygen supply is correlative with the level of oxygen saturation of hemoglobin and blood flow velocity. Together with pulse frequency, conceptual equation is established as follows:

H_eat=F(G_LU, B_V, P_OS, P_F)                       (2)

where  H_eat is the metabolic heat of human body; B_V is the blood flow velocity; P_OS is the oxygen saturation of hemoglobin; P_F is the pulse frequency; and G_LU is the blood glucose level.

Then, the function is denoted as follows:

G_LU=F(H_eat, B_V, P_OS, P_F)                       (3)

G_LU is calculated by the physiological parameters such as the metabolic heat, blood flow velocity, oxygen saturation of hemoglobin and pulse frequency.

2.2 Algorithm

Applying   the hydromechanical theory, the mathematical model of human convection coefficient was established. Metabolic rate in human fingers can be calculated based on the mathematical model of human heat balance. Using the improved heat elimination method, blood flow rate of human finger can be gained. As mentioned above, according to the conservation of energy method, the blood glucose level has a relationship with human metabolic rate, blood flow, oxygen saturation of hemoglobin and pulse frequency. According to the mathematical model established with multi-dimensional nonlinear regression analysis, BGL can be calculated. The relationship can be described as follows:

      (4)

where   M is metabolic rate,   is the average of M with the number of N, the same to ,and .

The relationship between M_i, B_Vi, P_OSi, P_Fi calculated above and G_LU (the blood glucose level detected with an automatic biochemical analyzer (ABA)) is established as follows:

G_LU=β×X                                   (5)

where

Using the well-known identities of the inverse of a partitioned matrix, the equation is gained as follows:

β=G_LU×X^-1                                 (6)

With Eq.(4) and the known β, the equation is as follows:

   (7)

where  ξ is defined as linear compensation parameter. ξ=-100 mg/L and ，，，， can be calculated with the Matlab software.

3 Non-invasive blood glucose sensor

3.1 Sensor selection

According to the system requirements, LM series, HIH3610 series, A2TPMI series, DLED-660/905 and magnesium sheet (2 mm×5 mm×50 mm) were chosen as temperature sensor, humidity sensor, radiation sensor, infrared sensor and metal sheet, respectively.

3.2 Non-invasive blood glucose sensor module

The sensor module made of Perspex is given in Fig.1. It involves sensors of temperature, humidity, radiation, infrared and magnesium sheet.

4 System hardware

NBGMA consists of two modules. The first is non-invasive glucose sensor module, and the second is detecting circuit board, which includes signal processing circuit, analog filter, amplifier, serial port communication circuit, LCD interface circuit, power supply. Blood oxygen module was designed for detecting oxygen saturation of hemoglobin and pulse frequency. System hardware block diagram is given in Fig.2.

Fig.1 Non-invasive blood glucose sensor module: 1—Environ- mental humidity sensor; 2—Environmental temperature sensor; 3—Temperature sensor for far-end of sheet metal; 4—Metal sheet; 5—Temperature sensor for near-end of sheet metal; 6—Finger surface humidity sensor; 7—Finger surface temperature sensor that is directly contacted with finger in measurement;  8—Infrared reception sensor; 9—Infrared radiation sensor;  10—Interface of power supply; 11—Interface of data bus; 12—Radiation sensor

Fig.2 System hardware block diagram

5 System software

Two important modules in the system software are non-invasive blood glucose apparatus software module and personal computer (PC) software module. The software module of NBGMA consists of three modules: data-gathering module, data-processing module and serial communication module. PC software module consists of two parts: data gathering module and database module. There are four main functions of data gathering module: serial communications, data processing, data acquisition and dynamic display, and data storage. Database module is responsible for data- searching, data-revising and data-displaying.

6 Experiments

6.1 Experimental setup

Schematics of experimental setup used in clinical study is illustrated in Fig.3. NBGMA receives the data from eight sensors for further processing. Then, the processed data together with the pulse frequency and oxygen saturation hemoglobin are sent to PC through serial port.

6.2 Subjects and method

A total of 20 healthy volunteers (12 males and 8 females) were involved in the experiment. Before measuring, tight control of environment was needed. The temperature and humidity of the room were almost constant. Additionally, during the experiment, volunteers were required to keep calm and to keep the finger clean, with no strenuous exercise and fasting. The experiment procedures are listed as follows.

Step 1 Keep the volunteer calm and the finger clean.

Fig.3 Schematics of experimental setup used in clinical study

Step 2 Sample venous blood (2 mL) and measure BGL by ANA (AUTOLAB18).

Step 3 Put the forefinger of right hand on the sensor module. 5-10 s later, start NBGMA.

Step 4 0.5 min later, remove the forefinger of right hand.

Step 5 Measurement was finished.

6.3 Results and discussion

Table 1 lists the results of clinical experiment by the non-invasive method against biochemical analysis for venous-blood samples as the reference method. The BGL ranges measured by NBGMA and ABA are 770-1 100 mg/L, 680-1 100 mg/L, respectively, which shows that two sets of data have a similar dynamic range. The means of 20 data from NBGMA and ABA are close to each other, and the difference of standard deviation (SD) is 24.7 mg/L. From statistics of the difference (|ε|) of each result, There are 19 sets of data are less than or equal to the margin of 100 mg/L, only one is more than 100 mg/L.

Fig.4 shows a regression analysis involving 20 data points. Each point represents that the non-invasive measurement and the venous blood collection for glucose oxidase measurement were simultaneously performed at random timing for each volunteer. ρ_BGL-NBGMA is the blood glucose concentration detected by NBGMA. ρ_BGL-ABA is the blood glucose concentration measured by ABA. The correlative coefficient between the blood glucose level by the two apparatuses is 0.807. Fig.4 also shows the clarke error grid analysis (EGA) which is accepted as one of “gold standards” for determining the accuracy of blood glucose meters. Values in regions A and B represent accurate and acceptable glucose results, respectively. As shown in Fig.4, almost all data are included in region A.

Reproducibility was measured at six intervals over 30 min for fasting volunteers without diabetes. The coefficient of variation (CV) was 4% at the mean BGL of 921.6 mg/L. Because relatively stable BGL could be obtained only from fasting volunteers without diabetes, it is necessary to measure reproducibility for healthy volunteers.

However, the correlation between the BGL by the two apparatuses is not adequate. Many factors affect the results. Firstly, the factor of measuring environment: in theory, the volunteer should be measured in condition of constant temperature, however this was not very tightly controlled in the clinical experiment. Secondly, the factor of experiment subjects: most of the volunteers could not remain clam throughout the measurement, which caused the motion artifact. Thirdly, the factor of the sensor module: in the clinical experiment, the temperature of the near-end and far-end of the metal sheet detected did not
Table 1 Results of clinical measurement from NBGMA and ABA

Note: ρ_BGL-ABA denotes BGL measured by automatic biochemical analyzer; ρ_BGL-NBGMA denotes BGL measured by non-invasive blood glucose measuring apparatus; |ε| denotes difference of BGL from ABA and NBGMA.

exactly satisfy the need in theory. At last, more factors such as the roughness of skin, age and sex should be involved in the algorithm. The results cannot be used to determine that BGL can be measured with the COEM precisely, but suggest that it is feasible.

Fig.4 Regression analysis on Clarke error grid involving 20 data points by non-invasive method against biochemical method for venous-blood samples as reference method (Region A: 100%, Regions B, C, D and E: 0%, n means number of measurements, and r means coefficient of correlation)

7 Conclusions

(1) COEM is developed for non-invasive measurement of blood glucose. The novel NBGMA is designed with COEM and the detecting algorithm based on COEM is presented.

(2) Clinical experiment is set up to test the accuracy of NBGMA. The BGL measured by ABA (AUTOLAB18) is considered as standard value. The result shows that the BGL measured by NBGMA is very close to that by ABA, and the correlation coefficient is 0.807. Reproducibility is measured. The CV is 4%. The results of Clarke error grid analysis (EGA) also indicate that it is feasible to measure blood glucose level with the conservation of energy method.

(3) This work is in progress. In the further study, the performance of the sensor module will be improved by using high-performance sensors, and the control of measuring environment will be strengthened. More clinical experiments, in which the diabetic will be involved, will be introduced for algorithm amendment. It is predicted that the correlation coefficient will increase.

[1] YANG Guang-huang. Pathology [M]. Beijing: People’s Medical Publishing House, 2004. (in Chinese)

[2] CHEN Zhen-cheng, JIN Xing-liang, XU Xiao-wen, ZHU Jian-ming, WANG Di-ya. A primary research of non-invasive glucose measuring apparatus[J]. Chinese Journal of Sensors and Actuators, 2008, 21(7): 1119-1123. (in Chinese)

[3] YAMAGUCHI M, KAWABATA Y, KAMBE S, WARDELL K, MYSTROM F H, NAITOH K, YOSHIDA H. Non-invasive monitoring of gingival crevicular fluid for estimation of blood glucose level[J]. Medical and Biological Engineering, and Computing, 2004, 43(3): 322-327.

[4] SUN Fei, KONG De-yi, MEI Tao. Technology and prospects of noninvasive blood glucose measurement[J]. Journal of Biomedical Engineering, 2005, 22(1): 171-174. (in Chinese)

[5] WEIJMAN I, ROS W J, RUTTEN G E, SCHAUFELI W B, SCHABRACQ M J, WINNUBST J A. Frequency and perceived burden of diabetes self-management activities in employees with insulin-treated diabetes: Relationships with health outcomes[J]. Diabetes Res Clin Pract, 2005, 68(1): 56-64.

[6] CHEN Zhen-cheng, FANG Cheng, WANG Hong-yan. A novel disposable amperometric UA strip[J]. Sensors and Actuators, B: Chemical, 2008, 129(2): 710-715.

[7] YU S Y, CHIU J H, YANG S D, HSU Y C, LUI W Y, WU C W. Biological effect of far-infrared therapy on increasing skin microcirculation in rats[J]. Photodermatology, Photoimmunology & Photomedicine, 2006, 22(2): 78-86.

[8] WEISS R, YEGORCHIKOV Y, SHUSTERMAN A, RAZ I. Noninvasive continuous glucose monitoring using photoacoustic technology-results from the first 62 subjects[J]. Diabetes Technol Ther, 2007, 9(1): 68-74.

[9] CHEN Zhen-cheng, WANG Hong-yan, HE Ji-shan, DENG Zhen-sheng. Development and clinical application of biosensor for detecting uric acid in whole-blood[J]. Chinese Journal of Analytical Chemistry, 2007, 35(4): 615-618. (in Chinese)

[10] SIEG A, GUY R H, GADO-CHARRO M B. Noninvasive and minimally invasive methods for transdermal glucose monitoring[J]. Diabetes Technol Ther, 2005, 7(1): 174-197.

[11] YAMAGUCHI M, KAMBE S, YAMAZAKI K, KOBAYASHI M. Error grid analysis of noninvasive glucose monitoring via gingival crevicular fluid[J]. IEEE Trans Biomed Eng, 2005, 52(10): 1796- 1798.

[12] CHEN Zhen-cheng, GONG Qiong, WANG Di-ya, XU Xiao-wen, ZHU Jian-ming, JIN Xing-liang. Design of non-invasive blood sugar measuring sensor based on energy metabolic conservation[J]. Chinese Journal of Sensors and Actuators, 2008, 21(5): 724-728. (in Chinese)

[13] KUWA K. Noninvasive blood glucose monitoring: New technology using metabolic heat conformation method[J]. Rinsho Byori, 2006, 54(5): 519-525.

[14] CADUFF A, HIRT E, FELDMAN Y, HEINEMANN L. First human experiments with a novel non-invasive, non-optical continuous glucose monitoring system[J]. Biosens Bioelectron, 2003, 19(3): 209-217.

[15] CHEN Wen-liang, CUI Hou-xin, LIU Rong, XU Ke-xin, CHEN Min-sen. Experimental research on OGTT for non-invasive blood glucose detection through near-infrared spectroscopy ranging from 1 100 nm to 1 700 nm[J]. Journal of Biomedical Engineering, 2004, 21(5): 824-827. (in Chinese)

[16] CHO O K, KIM Y O. Blood sugar level measuring apparatus: US, 2004/0225209[P]. 2004-11-11.

(Edited by YANG You-ping)

Foundation item: Project(07JJ6133) supported by the Natural Science Foundation of Hunan Province, China

Received date: 2008-12-08; Accepted date: 2009-03-02

Corresponding author: CHEN Zhen-cheng, Professor, PhD; Tel: +86-731-88877053; E-mail: chzhch@mail.csu.edu.cn