Applied Mathematical Sciences, Vol. 5, 2011, no.

6, 279 - 286

Mathematical Model for Detecting Diabetes in the Blood
B. Kwach Department of Mathematics and Applied Statistics Maseno University, Box 333, Maseno, Kenya O. Ongati Department of Mathematics and Applied Statistics Maseno University, Box 333, Maseno, Kenya R. Simwa School of Mathematics University of Nairobi, Box 30197, 00100, Nairobi, Kenya
Abstract This study presents a new mathematical model for Blood Glucose Regulatory System(BGRS) which includes epinephrine as a third vari˙ = AY, and whose solution has been analyzed for able in the form, Y equilibrium and stability to provide the blood glucose concentrations for diabetics and non-diabetics. We establish that the final model is asymptotically stable compared to the existing models, that is, the eigenvalues of the coefficient matrix are complex numbers with negative real parts. Furthermore, the resonance period for the final model, that is, T0 = 2.9847134 hours, is far less than that of the existing model, showing that the glucose concentration returns to normal level within a shorter time.

Mathematics Subject Classification: Primary 93A30; Secondary 91B74, 93C05 Keywords: Mathematical model, Linear system, Resonance period

O. Ongati and R. Middleman [21]. there are over 20 million diabetics in America. P. Duff and Jason [13]. (a) increased glucose production from glycogen breakdown (b) increased glucose production from lactate and amino acids (c) increased fat mobilization by stimulation of hormone sensitive lipase (d) small net stimulation of insulin secretion from pancreatic β -cells . Central. After the ingestion of the glucose load a serious shortcoming of this simplified model was experienced because variables such as epinephrine and glucagon. Nairobi.3 million Kenyans suffer from diabetes. including Intra Venous Glucose Tolerance Test (IVGTT). Eastern and Rift Valley provinces. The majority of mathematical models were devoted to the dynamics of glucose-insulin. relating important public health questions to basic infection parameters.280 B. all the existing models were based on two variables only: glucose and insulin. Glucose plays an important role in the food metabolism of any vertebrate tissue since it is a source of energy for all tissues and organs. Diabetes Mellitus is a disease which is characterized by too high sugar levels in the blood and urine. Every 2 years 921 new cases are diagnosed in various clinics in Nairobi. Kenya that 3. include. So far. Epinephrine is secreted by the adrenal medulla in response to acute stress (fight or flight response). Cases of diabetes in the country have increased from 3. Coast. During the next three to five hours several measurements are made on the concentration of glucose in the patient’s blood and these measurements are used in the diagnosis of diabetes Mellitus. some of which are highlighted in the appendix. Kwach. In the GTT. Diabetic patients require supplement of insulin in the form of regular injections and tablets in addition to modified diet to regulate glucose input. It was established by the Kenya Diabetes Management and Information Center during the free diabetes screening exercise at M. Nyanza. It is quite conceivable. Reporter [25]. Okwemba [22]. Oral Glucose Tolerance Test (OGTT) and Frequently Sampled Intra Venous Glucose Tolerance Test (FSIVGTT). It is usually diagnosed by means of a glucose tolerance test (GTT). which play an important role during this time of recovery phase of the GTT response were neglected. Krimmel et al [20]. therefore that the body will interpret this as an extreme emergency and thereafter the hormones epinephrine and glucagon come in play. Today. Simwa 1 Introduction Over the past years.5 to 10 per cent of the population in the past one year. Shah hospital. mathematics has been used to understand and predict the spread of diseases. an individual comes to the hospital after an overnight fast and is given a large dose of glucose (sugar in the form in which it usually appears in the bloodstream). Important effects of epinephrine. six million of whom must take injections of insulin daily. Eastham [14].

k . c. then at equilibrium. f . Thus.r. Thus. d. It is assumed that the rate of change of these quantities depend only on the values of g and h. If either h or g is a non-zero value then the body tries to restore the equilibrium. epinephrine is included as a separate variable in this model of blood glucose regulatory system. l. If there is an internal rate at which the blood glucose concentration is being increased. a. if it is assumed that there is no recent digestion. dg dh de d2 g = −a − b + f 2 dt dt dt dt Substituting for dh = cg − dh + ke dt and de = −lg − mh + ne dt (2) . (1) 3 Mathematical Model for Testing Diabetes dg = −ag − bh + f e dt From the model in equation (1). the following systems of differential equations results. differentiating of equation (1) w. g ˙ = −ag − bh + f e ˙ h = cg − dh + ke e ˙ = −lg − mh + ne where e represents epinephrine. If g is taken to be excess glucose concentration and h is excess insulin concentration at time t. m and n are constants. we obtain.t t. g = h = 0. glucose and insulin concentration will be in equilibrium.Mathematical model for detecting diabetes in the blood 281 2 Preliminaries Provided there is no recent digestion. Ackerman et al [1]. b. positive value of g or h corresponds to concentrations greater than the equilibrium values while negative values corresponds to concentrations less than equilibrium values.

O. This means that characteristic equation of (2) has complex roots.78 0. (4) 4 Verification of the model      g ˙ −2.24 g ˙  =  0.282 of equation (1) in equation (2) gives. which is the system natural frequency is the basic descriptor of the response to a GTT.34 1. Kwach. B. (5) . 2 If α2 − ω0 > 0.92 −4. and assuming h = 0 d2 g bk dg bka + ( + a − n) + (bc + f l + − na)g = 0 2 dt f dt f This is in the form of. then g (t) drops very rapidly from a fairly high values to negative ones below the equilibrium value.98 0. d2 g dg 2 + 2 α + ω0 g=0 dt2 dt 1 bk ⇒ α = ( + a − n) 2 f and 1 2 ω0 = ( bka + bc + f l − na) f where the value of ω0 . where e= 1 dg ( + ag ) f dt (3) from equation (1). Simwa d2 g dg + a + (bc + f l)g + (bk − f n)e + (f m − bd)h = 0 dt2 dt From equation (3). the model is in the form of. Ongati and R.53 e e ˙ From equation (1). so α2 − ω0 < 0.14   h   h −2. The model certainly conforms to reality in predicting that the blood glucose concentration tends to return eventually to its optimal concentration (1).208 −0. It is 2 2 assumed that α2 − ω0 is negative.94 −0. The body will interpret this as an extreme emergency and large amounts of epinephrine will be secreted (2).

0050.208. it was shown that by employing the model.24.7 2.4745 6.10 7.0831 7.34.20 4.9385 2.78. in mmol/L.94. 4.9766 Glucose Insulin Initial Control 30.20 Insulin Epinephrine Control Control 6.5476). 2 ⇒ ω0 = 4. This is a situation where the individual is normal.0 2. From the relation.43152 ⇒ ω0 = 2. the individual is said to be normal.2616 For individuals presenting with very low blood sugar.2 27.5910 3.92.98 and 0.00 17.5754i.0 3. that is hypoglycaemic patients. secondary values and raw data of blood glucose concentrations before .60 7.3 20. 0.1 18.00 25.9 2.53 respectively are secondary values. Since the eigenvalues are complex numbers with negative real parts.70 7. Table (3.8 3.9847134 Since the resonance period is far less than 4 hours.5910 2.00 9. 2π T0 = ω0 and by substituting ω0 .0 3.90272 + 3.Mathematical model for detecting diabetes in the blood 283 where the constants given by 2.6456 − 1. 1 2 ω0 = ( bka + bc + f l − na) f 2 ω0 = (1.00 18. the system is asymptotically stable. 1.9789 2.2 Epinephrine Control 3.5825 + 1. 0.4308 + 0.6415 5.7898 8.0634 11. Y Using MATLAB.2156 2. From the relation. 2π T0 = 2. Paolo et al [24].5825 − 1.1051176. This system of equations can be expressed in the matrix vector form as ˙ = AY.9093 8. the blood glucose concentrations were corrected to normal levels in hyperglycaemic situations.6495 3.1): Blood glucose concentrations after Insulin infusion and corresponding concentrations with Epinephrine.7 18. 2.30 22. 0. λ1 = −1.30 18.84 7. Glucose Initial 26.46 6.0 18. 0.5754i and λ3 = −0. λ2 = −1. the eigenvalues are.14.3 7. From the results.1051176 ⇒ T0 = 2.9 21.00 20.

Epinephrine has been successfully incorporated as a third variable in this model of blood glucose regulatory system (BGRS).284 B.2 1.1 4.5 2. Ongati and R. the blood glucose concentrations were corrected to normal levels in hypoglycaemic states.1 5. The importance of this third variable lies in its ability to help in conducting a reliable test for detecting diabetes in the blood. Furthermore the resonance period for this model which is T0 = 2. This leads to a system of ˙ = AY and linear homogenous equations.1 1.0 2.3 5.8 1.8 5. Table (3.2 Similarly from the results.1 4. Fasting blood sugar(mmol/L) 2.0 7. it is shown that by employing the model.9847134 hrs.2) below.0 2.8 5. Simwa and after intravenous (IV) bolus of 50 percent dextrose were used to correct the blood glucose concentrations as shown in table (3.7 1. This is exemplified by the sample results considered in Table (3. 5 Conclusion and recommendation This paper presents a model for detecting diabetes Mellitus in the blood described by equation (5). This model has been found to be asymptotically stable since the eigenvalues of the coefficient matrix are complex numbers with negative real parts.5232581 hrs for the existing model.2): Blood glucose concentrations before and after intravenous (IV) bolus of 50 percent dextrose. Kwach.6 6. which are expressed in the form Y whose solution provides the blood glucose concentrations for diabetics and non diabetics.2).3 1. is far less than T0 = 3.9 1.6 4.1 Corrected blood sugar(mmol/L) 6. Future research may take into consideration an external rate at which the blood glucose concentration is being increased. This shows that the glucose concentration returns to normal level within a shorter time. It is worth noting that the model developed in this study only considered an internal rate at which the blood glucose concentration is being increased. O.1) and Table (3. .

Adrenaline. New York. G. 325-336. Marcel Dekker. Brooks/Cole Publishing Company. Concepts and Models of Biomathematics. Differential Equations. (2002). Houston K. [3] Berry J. [4] Blanchard P. C. Endocrine Society. Boutayeb A. 783. Differential Equations and Their Application: An Introduction to Applied mathematics. [7] Burge R. Ider Y. Ordinary Differential Equations and Stability Theory. Cobelli C. W. J W Arrowsmith Ltd. (1993). G. Mark. References [1] Ackerman E. The Journal of Clinical Endocrinology and Metabolism. Fallon S. [PubMed]. Thomaseth K. Mathematical Modelling of the Intravenous Glucose Tolerance Test. S. The Fourfold Path to Healing. The minimal model of glucose disappearence : optimal input studies. New Mexico. 23: E667-E677. Devaney L. 40:136-168. NewTrends Publishing. 7th Edition. 131-156. [9] Cowan T. (2000).(1996). [8] Cobelli C. Quantitative Estimation of Insulin Sensitivity. 35:911-917.Mathematical model for detecting diabetes in the blood 285 Acknowledgements: We thank Doctor Amos Otedo for his useful comments on one of our discussions and allowing us to access the secondary data. [6] Brawn M. Taher A. Am J Physiol. McMillan J (2004). [5] Bolie V. Bristol. (1960). J Math Biol [PubMed]. Coefficients of Normal Glucose Regulation. Rosevar J. (1987). Qualis R. Sobhy. Journal of Applied Physiology. Raghavendra V. Jason. [13] Duff. [2] Bergman R. 16. N. Heinmets. Mathematical Modelling. Arino O. Survey of Organic Chemistry Molecule Projects. [10] Derouich M. D. R. Math Biosciences [PubMed]. (1995). [12] Deo S. (2000). [11] De Gaetano A. F. (1975). 83:127-130. The effect of physical exercise on the dynamics of glucose and insulin [PubMed]. (1969). W. (2001). . (1979). Molnar G. Hall R. Springer. Schade S.

Diabetes Health Centre. Simwa [14] Eastham R. Drug Infusion Control: An Extended Direct Model Reference Adaptive Control Strategy. Counter Current Oxygen Exchange in The Swim Bladder of DeepSea Fish. M. (2003). New York. Fundamentals of Clinical Endocrinology. Ellis Horwood Ltd. says WHO . 45-60. M (1983). Ongati and R. Mathematical and Computer Modelling. (2002). Besser M (1974). D. [26] Rodi E.286 B. [19] King W. 8/7/2007). WebMD Medical Reference. [21] Middleman. Health Crisis.Willey Interscience. [22] Okweba A. (1997). (1989). 12. (1992). Am J Physiol Endocrinol Metab 283: E78-E84. Linear and Non-Linear Differential Equations. [20] Krimmel E. [24] Paolo V. Johnson R. [15] Ebadi M. 2007). ELBS and Pitman Medical Publishing Company. Kwach. Epinephrine effects on insulin-glucose dynamics. (14th November. Spilker M E. 389. [16] Hall R. Core Concepts in Pharmacology. 2010 . [18] Huntley I. (2007). New York. 2nd Edition. D (1983). Diabetes Cases on the rise. 67-69. 14/6/2007). Received: July. Lippincott-Raven Publishers. Smart A. O. pp. (Sunday. Anderson J. Bristol BS4 5NU. The Low Blood Sugar Handbook. Nation Media Group. Transport Phenomena in the Cardiovascular System. Nation Media Group. [17] Healthwise. John Wright and Sons Ltd. (Thurday. Medical Biochemistry [PubMed]. 5th Edition. Y and Jacques’s. Franklin Publishers. New York. Gallo A. pp 5. A Laboratory Guide to Clinical Diagnosis. [23] Palerm CCR. Krimmel P. [25] Reporter. Cobelli C. Avogaro A. (1972). G.

Sign up to vote on this title
UsefulNot useful