Methods for improving performance and reliability of biosensors

• US 6,885,883 B2
• Filed: 05/14/2001
• Issued: 04/26/2005
• Est. Priority Date: 05/16/2000
• Status: Active Grant

• Current Assignee: LifeScan IP Holdings, LLC
• Original Assignee: Cygnus, Inc. (Johnson & Johnson)
• Inventors: Potts, Russell O., Tierney, Michael J., Parris, Norman A., Uhegbu, Christopher
• Primary Examiner(s): Dahbour, Fadi H.
• Application Number:
  US09/859,218
• Publication Number:
  US 20020026110A1
• Time in Patent Office: 1,443 Days
• Field of Search: 600/345, 600/346, 600/347, 600/348, 600/349, 600/350, 600/365, 600/322, 600/368, 600/576, 204/403.01, 128/898, 435/14, 435/817, 604/20
• US Class Current: 600/347
• CPC Class Codes: A61B 5/14514 using means for aiding extraction of interstitial fluid, e.g. microneedles or suction
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/1486 using enzyme electrodes, e.g. with immobilised oxidase

Abstract

The present invention relates to a predictive-kinetic method for use with data processing of a sensor-generated signal, as well as, microprocessors and monitoring systems employing such a predictive-kinetic method. Data from a transient region of a signal is used with suitable models and curve-fitting methods to predict the signal that would be measured for the system at the completion of the reaction. The values resulting from data processing of sensor response using the methods of the present invention are less sensitive to measurement variables.

Claims

1. One or more microprocessors, comprising programming to control

(i) operating a sensing device for obtaining a measured signal over time, comprising a measured signal response curve, from a sample comprising glucose collected from a biological system, wherein said measured signal is specifically related to the amount or concentration of glucose present in the biological system, and said measured signal response curve comprises kinetic and equilibrium regions;

and

(ii) a computation method using (a) a mathematical model comprising selected parameters, wherein said model describes the measured signal response curve, and said mathematical model is selected from the group consisting of a first order process, combined first order and zero order process, a parallel multiple first order process, a flux process, and an n^th order process, and (b) an error minimization method to iteratively estimate values of the parameters using said model to fit a predicted response curve to said measured signal response curve, wherein (a′

) the error minimization method provides a calculated error based on differences between said predicted and measured signal response curves, and (b′

) said estimating is iteratively performed until the calculated error between the predicted and measured signal response curves falls within an acceptable range or until no further statistically significant change is seen in the calculated error, at which time iterative estimation of the parameters is stopped, said iterative estimation and error minimization results in a predicted response curve corresponding to said measured signal response curve, said predicted response curve yields a predicted end-point value and a measurement correlated to the amount or concentration of the glucose.

2. The one or more microprocessors of claim 1, further programmed (i) to operate said sensing device to obtain a series of measured signals over time resulting in a series of measured signal response curves, and (ii) to provide a predicted response curve corresponding to each measured signal response curve.

3. The one or more microprocessors of claim 2, wherein after estimation of each predicted response curve for each measured signal response curve in the series of measurements an amount or concentration of the glucose is determined based on the predicted response curve.

4. The one or more microprocessors of claim 2, wherein a first mathematical model is selected to estimate at least one predicted end-point value corresponding to a measured signal response curve and at least one different mathematical model is used to estimate at least one predicted end-point value corresponding to a different measured signal response curve.

5. The one or more microprocessors of claim 2, wherein a single mathematical model is selected to estimate the predicted end-point values for all measurements in the series.

6. The one or more microprocessors of claim 1, wherein said measured signal response curve comprises a measurement of current over time, or measurement of charge over time.

7. The one or more microprocessors of claim 1, wherein said measured signal response curve comprises a measurement of current over time, said predicted end-point value is an estimated signal at equilibrium, where said predicted end-point value provides a predicted final background value, and said measurement correlated to the amount or concentration of glucose corresponds to an area under the predicted response curve.

8. The one or more microprocessors of claim 7, wherein said area under the predicted response curve is obtained by integration of the predicted response curve.

9. The one or more microprocessors of claim 8, wherein before said integration is performed said final background value is used to perform a background subtraction correction of the predicted response curve and said measurement correlated to the amount or concentration of glucose corresponds to an area under the predicted response curve.

10. The one or more microprocessors of claim 9, wherein the end-point value of the integrated predicted response curve is converted to an amount or concentration of the glucose.

11. The one or more microprocessors of claim 8, wherein the end-point value of the integrated predicted response curve is converted to an amount or concentration of the glucose.

12. The one or more microprocessors of claim 11, wherein conversion of the end-point value of the integrated predicted response curve to an amount or concentration of glucose is performed by a method comprising applying a calibration value.

13. The one or more microprocessors of claim 7, wherein said mathematical model comprises a parallel multiple first order process.

14. The one or more microprocessors of claim 13, wherein said parallel multiple first order process comprises the following:

S_t=S₀+S₁*e<sup>−

k1*t</sup>+S₂*e<sup>−

k2*t</sup>+final_Bkgrd 

 

Eqn. 20

where S_o is response at t=0, t is time, S_t is a total signal at time t, S₁ and S₂ are signals at time t consistent with two processes associated with apparent rate constants k₁ and k_2, and final_Bkgrd is an estimated signal response at completion of a reaction used to obtain the measured signal.

15. The one or more microprocessors of claim 14, wherein the area under the predicted response curve is obtained by integration.

16. The one or more microprocessors of claim 15, wherein before said integration is performed said final_Bkgrd value is used to perform a background subtraction correction of the predicted response curve and said measurement correlated to the amount or concentration of glucose corresponds to the area under the predicted response curve.

17. The one or more microprocessors of claim 13, wherein said parallel multiple first order process comprises the following:

S_t=S_t*e<sup>−

kl*t</sup>+S₂*e<sup>−

k2*t</sup>+final_Bkgrd 

 

Eqn. 21

where t is time, S_t is a total signal at time t, S₁ and S₂ are signals at time t consistent with two processes associated with apparent rate constants k₁ and k₂, and final_Bkgrd is an estimated signal response at completion of a reaction used to obtain the measured signal.

18. The one or more microprocessors of claim 17, wherein the area under the predicted response curve is obtained by integration.

19. The one or more microprocessors of claim 18, wherein before said integration is performed said final_Bkgrd value is used to perform a background subtraction correction of the predicted response curve and said measurement correlated to the amount or concentration of glucose corresponds to the area under the predicted response curve.

20. The one or more microprocessors of claim 1, wherein said one or more microprocessors are further programmed to control operating a sampling device to provide said sample comprising glucose.

21. The one or more microprocessors of claim 20, wherein the sampling device comprises one or more collection reservoirs into which the sample is collected.

22. The one or more microprocessors of claim 21, wherein the sampling device comprises an inotophoretic device to collect the sample comprising glucose from the biological system into at least one collection reservoir.

23. The one or more microprocessors of claim 21, wherein the collection reservoir comprises an enzyme that reacts with the glucose to produce an electrochemically detectable signal.

24. The one or more microprocessors of claim 23, wherein the enzyme comprises glucose oxidase.

25. The one or more microprocessors of claim 20, wherein the sampling device comprises a laser device or a sonophoresis device.

26. A monitoring system comprising the one or more microprocessors of claim 1, wherein said monitoring system further comprises the sensing device.

27. The one or more microprocessors of claim 1, wherein the mathematical model further comprises a zero-order component.

28. The one or more microprocessors of claim 1, wherein said mathematical model comprises more than one process and each process comprises selected parameters.

29. The one or more microprocessors of claim 28, wherein each process has a corresponding weighting factor.

30. The one or more microprocessors of claim 1, wherein said measured signal response curve comprises data points.

31. The one or more microprocessors of claim 30, wherein at least three data points are obtained from the kinetic region of the measured signal response curve, and these data points are used to estimate the half-life of the measured signal.

32. The one or more microprocessors of claim 31, wherein said obtaining of the measured signal continues for a time period corresponding to at least three half-lives of the signal.

33. The one or more microprocessors of claim 31, wherein the estimate of the half-life (t_½

) further comprises, estimating a rate constant (k) for a first order model comprising

S_t=S_∞

−

(S_∞

−

S_o)e<sup>−

kt</sup>  

 

(Eqn.

     1)

where S_o, S_t, and S_∞

are initial, intermediate, and end-point signals, k and t are the observed first-order rate constant and time, respectively, wherein estimating said rate constant is performed by a method comprising plotting the natural log of signal (S_t−

S_o) over time, where the slope of the resulting line corresponds to an estimate of k, and the half-life of the signal is calculated by using t_½

=1n2/k.

34. The one or more microprocessors of claim 1, wherein said obtaining of the measured signal is carried out for a defined period of time.

35. The one or more microprocessors of claim 1, wherein before applying said computation method the measured signal response curve is integrated before using the mathematical model to fit the predicted signal response curve to the measured signal response curve.

36. The one or more microprocessors of claim 1, wherein said computation method further comprises integration of the predicted response curve after using the mathematical model to fit the predicted signal response curve to the measured signal response curve.

37. The one or more microprocessors of claim 1, wherein said mathematical model comprises a first order process.

38. The one or more microprocessors of claim 37, wherein said first order process comprises the following:

S_t=S_∞

−

(S_∞

−

S_o)e<sup>−

kt</sup>  

 

(Eqn.

     1)

where S_o, S_t, and S_∞

are initial, intermediate, and end-point signals, k and t are the observed first-order rate constant and time, respectively.

39. The one or more microprocessors of claim 1, wherein said mathematical model comprises a parallel multiple first order process.

40. The one or more microprocessors of claim 39, wherein said parallel multiple first order process comprises the following:

S_t=S_o+(S<sub>∞

1</sub>−

S_o)*(1−

e<sup>−

k1t</sup>)+(S<sub>∞

2</sub>−

S_o)*(1−

e<sup>−

k2t</sup>)+(S<sub>∞

3</sub>−

S_o)*(1−

e<sup>−

k3t</sup>)+ . . .  

 

(Eqn. 6A)

where S_o, S_t are initial and intermediate signals, S<sub>∞

1</sub>, S<sub>∞

2</sub>, S<sub>∞

3</sub>, etc., are end-point signals (related to k₁, k₂, k₃, etc., respectively), k₁, k₂, k₃, etc., are the observed first-order rate constants, and t is time.

41. The one or more microprocessors of claim 40, wherein the predicted end-point value is described by the following equation

S_∞

=(S<sub>∞

1</sub>+S<sub>∞

2</sub>+S<sub>∞

3</sub>+ . . . )+S_o 

 

(Eqn. 6B)

42. The one or more microprocessors of claim 40, wherein a change in the predicted end-point value relative to the initial signal is described by the following equation

Δ

S_∞

=(S<sub>∞

1</sub>+S<sub>∞

2</sub>+S<sub>∞

3</sub>+ . . . )  

 

(Eqn. 6C)

43. The one or more microprocessors of claim 39, wherein said parallel multiple first order process comprises the following:

S_t=S_o+(S<sub>∞

1</sub>−

S_o)*(1−

e<sup>−

k1t</sup>)+(S<sub>∞

2</sub>−

S_o)*(1−

e<sup>−

k2t</sup>)  

 

(Eqn.

     10)

where S_o, and S_t, are initial and intermediate signals, S<sub>∞

1</sub>, and S<sub>∞

2</sub> are end-point signals (related to k₁ and k₂, respectively), k₁, k₂, and t are the observed first-order rate constants and time.

44. The one or more microprocessors of claim 43, wherein said parallel multiple first order process further comprises at least one zero order process.

45. The one or more microprocessors of claim 44, wherein said parallel multiple first order process comprises the following:

S_t=S_o+k_ot+(S<sub>∞

1</sub>−

S_o)*(1−

e<sup>−

k1t</sup>)+(S<sub>∞

2</sub>−

S_o)*(1−

e<sup>−

k2t</sup>)+(S<sub>∞

3</sub>−

S_o)*(1−

e<sup>−

k3t</sup>)  

 

(Eqn.

     16)

where S_o, S_t are initial and intermediate signals, S<sub>∞

1</sub>, S<sub>∞

2</sub>, S<sub>∞

3</sub>, are end-point signals (related to k₁, k₂, k₃, respectively), k₁, k₂, k₃, are the observed first-order rate constants, k_o is a zero order rate constant, and t is time.

46. The one or more microprocessors of claim 39, wherein said parallel multiple first order process further comprises at least one quadratic or square root term.

47. The one or more microprocessors of claim 1, wherein said mathematical model comprises an n^th order process.

48. The one or more microprocessors of claim 47, wherein said n^th order process comprises the following:

S_t=S_∞

(−

/+)[k(n−

1)*t(+/−

) (S_∞

−

S_o)<sup>1−

n</sup>]<sup>1/(1−

o)</sup> 

 

(Eqn.

     8)

where S_o, S_t, and S_∞

are initial, intermediate, and end-point signals, k and t are the observed rate constant and time, n is the order of the process, where n does not equal 1, and for (−

/+) the first function (−

) is used for data that increase in magnitude as a function of time, and the second function (+) is used for the reverse case, correspondingly for (+/−

) the first function (+) is used for data that increase in magnitude as a function of time, and the second function (−

) is used for the reverse case.

49. The one or more microprocessors of claim 1, wherein said mathematical model comprises a flux model.

50. The one or more microprocessors of claim 49, wherein said flux model comprises the following:

$\begin{array}{cc}{S}_t=S_o+\left(S_{\mathrm{\infty <br><br>}}-S_o\right)<br><br>\left[1+2<br><br>\underset{i=0}{\overset{\mathrm{\infty <br><br>}}{\sum <br><br>}}<br><br>{\left(-1\right)}^i<br><br>\exp <br><br>\left(-k_i<br><br>t\right)\right]& \text{(Eqn. <br><br> <br><br>9)}\end{array}$

where S_o, S_t, and S_∞

are initial, intermediate, and final (or end-point) signals, k_j=ki²π

², k is the characteristic diffusion rate constant, t is time, and i is a dummy-variable.

51. The one or more microprocessors of claim 1, wherein said sensing device employs electrochemical detection of glucose.

52. The one or more microprocessors of claim 51, wherein said electrochemical detection comprises the use of glucose oxidase.

53. A monitoring system for frequent measurement of glucose amount or concentration present in a subject, said system comprising, in operative combination:

(A) a sensing device for obtaining a measured signal over time, comprising a measured signal response curve, from a sample comprising glucose collected from the subject, wherein said measured signal is specifically related to the amount or concentration of glucose present in the subject, and said measured signal response curve comprises kinetic and equilibrium regions;

(B) one or more microprocessor(s) in operative communication with the sensing device, wherein said microprocessor is capable of (i) controlling the sensing device to obtain a series of measured signals, in the form of measured signal response curves, at selected time intervals, (ii) predicting measurement values for each measured signal in the series by employing (a) a mathematical model comprising selected parameters, wherein said model describes the measured signal response curve, and said mathematical model is selected from the group consisting of a first order process, combined first order and zero order process, a parallel multiple first order process, a flux process, and an n^th order process, and (b) an error minimization method to iteratively estimate values of the parameters using said model to fit a predicted response curve to said measured signal response curve, wherein (a′

) the error minimization method provides a calculated error based on differences between said predicted and measured signal response curves, and (b′

) said estimating is iteratively performed until the calculated error between the predicted and measured signal response curves falls within an acceptable range or until no further statistically significant change is seen in the calculated error, at which time iterative estimation of the parameters is stopped, said iterative estimation and error minimization results in a predicted response curve corresponding to said measured signal response curve, said predicted response curve yields a predicted end-point value and a measurement correlated to the amount or concentration of the glucose, and (iii) converting each measurement correlated to the amount or concentration of the glucose in the series to a measurement value indicative of the amount or concentration of glucose present in the subject.

54. The monitoring system of claim 53, wherein said measured signal response curve comprises a measurement of current over time, or measurement of charge over time.

55. the monitoring system of claim 54, wherein said measured signal response curve comprises a measurement of current over time, said predicted end-point value is an estimated signal at equilibrium, where said predicted end-point value provides a predicted final background value, and said measurement correlated to the amount or concentration of glucose corresponds to an area under the predicted response curve.

56. The monitoring system of claim 55, wherein said area under the predicted response curve is obtained by integration of the predicted response curve.

57. The monitoring system of claim 56, wherein before said integration is preformed said final background value is used to perform a background subtraction correction of the predicted response curve and said measurement correlated to the amount or concentration of glucose corresponds to an area under the predicted response curve.

58. The monitoring system of claim 57, wherein the end-point value of the integrated predicted response curve is converted to an amount or concentration of the glucose.

59. The monitoring system of claim 56, wherein the end-point value of the integrated predicting response curve is converted to an amount or concentration of the glucose.

60. The monitoring system of claim 53, further comprising a sampling device.

61. The monitoring system of claim 60, wherein said one or more microprocessors are further programmed to control operating the sampling device to provide said sample comprising glucose.

62. The monitoring system of claim 61, wherein the sampling device comprises one or more collection reservoirs into which the sample is collected.

63. The monitoring system of claim 62, wherein the sampling device comprises an iontophoretic device to collect the sample comprising glucose from the subject into at least one collection reservoir.

64. The monitoring system of claim 63, wherein the collection reservoir comprises an enzyme that reacts with the glucose to produce an electrochemically detectable signal.

65. The monitoring system of claim 64, wherein the enzyme comprises glucose oxidase.

66. The monitoring system of claim 60, wherein the sampling device comprises a laser device or a sonophoresis device.