GB2397375A

GB2397375A - Measuring analyte concentration in a fluid sample by illuminating the sample at two wavelengths

Info

Publication number: GB2397375A
Application number: GB0300764A
Authority: GB
Inventors: Gabor Tajnafoi
Original assignee: Hypoguard UK Ltd
Current assignee: Arkray Factory Ltd
Priority date: 2003-01-14
Filing date: 2003-01-14
Publication date: 2004-07-21
Also published as: EP1583841A1; WO2004063393A1; GB0300764D0

Abstract

The apparatus (100) comprises a first optical wavelength source (31), a second wavelength source (32), an optical detector (40), and an opening (22). The first wavelength source and second wavelength source are aligned to illuminate a fluid sample (15) received within the opening. The optical detector is aligned to measure the light reflected from the fluid sample at the first and second optical wavelengths. A control means (50) determines the analyte concentration in accordance with the reflected light measured at the said wavelengths. The said sources and the detector may be parallel, or there may be a rotational offset ( of e.g. 55 to 70 degrees) between the detector and the sources. The said first and second optical sources may have wavelengths of 700nm and 575nm respectively.

Description

DEVICE AND METHOD

The present invention relates to a device that is used to measure changes in colour chemistry in order to quantitatively estimate the concentration of a constituent chemical in a solution, and a method of operating such a device.

There is an increasing demand for highly portable medical diagnostic equipment that is user-friendly permitting operation by patients without intervention from medically trained staff. Such diagnostic equipment enables patients to monitor the concentration of a chemical constituent in biological material, for example glucose concentration in blood, and to seek further medical advice in the event that the measured value exceeds a threshold or is outside of a predetermined range of values. It can also be used by medically trained staff, at the bedside, in a hospital or in a nursing home.

Currently available blood glucose reflectance technology consists of a hand held reflectance meter and a disposable chemistry containing strip. Blood is applied to a sample application area on the uppermost surface of the strip. The strip consists of a support, for easy handling, and at the sample application area specialized structures which contain chemistries for measuring glucose as well as removing cellular material from the blood. Colour is generated at the lower surface of the strip which is proportional to the glucose concentration.

Conventional systems operate by applying blood to a strip which has been placed within a meter, in-meter dosing.

There are a number of incidences, particularly in nursing homes or in patients with impaired vision where it is preferable to apply blood to the strip external to the meter and then place the strip within the meter for measurement, off-meter dosage. It would therefore be desirable to have a system which can operate in both an in-meter and off-meter situation.

Reflectance meters must contain a light source. The light source can vary in intensity with time. Hence, it is important in conventional meters that the incident radiation from the light source and the resultant reflection from the unused test strip, described as feedback, is known. This is achieved in conventional meters by: a) Estimating the feedback at the time of production.

The tightness of the feedback will vary with the age of the meter. This method can also not compensate for batch-to-batch variations in the reflective surface of the test strip.

b) Incorporation of an internal reference surface or other light route to estimate the feedback.

c) A "relative" measurement between the initial unused colour of the strip to the colour developed stage.

It would be very difficult to do this in an off meter situation.

In all of the above configurations the time at which the - 3 - analytical reflectance measurement is made is very important to ensure optimal precision and clinical accuracy. This can be readily achieved in an on-meter dosing system but is prone to operator delays in an off- meter dosing system.

For these reasons, an alternative method for interrogating the analytical signal was sought which would allow on- meter and off-meter dosing with minimum operator involvement.

Some conventional diagnostic devices comprise a source for two or more optical wavelengths then one of the wavelengths is used as described above to determine the concentration of the target constituent chemical in the fluid sample whilst one or more further wavelengths are used for a different function, either to assist the calibration of the device or to reduce any unwanted optical effects that may be caused by effects other than the colour change that is due to the concentration of the target constituent chemical, for example contamination of part of the plastic film by the sample, such as blood cells.

According to a first aspect of the present invention there is provided an apparatus for measuring analyte concentration in a fluid sample, the apparatus comprising: a first optical wavelength source and a second optical wavelength source; an optical detector and an aperture, - 4 - the first optical wavelength source and the second optical wavelength source being aligned, in use, to illuminate a fluid sample received within the aperture, the optical detector being aligned, in use, to measure light reflected from the fluid sample at the first optical wavelength and the second optical wavelength, the control means determining the analyte concentration in accordance with the reflected light measured at the first optical wavelength and the reflected light measured at the second optical wavelength. Preferably the first optical wavelength source and the second optical wavelength source are substantially parallel and there is a rotational offset between the optical detector and the first optical wavelength source and the second optical wavelength source The rotational offset between the optical detector and the first optical wavelength source and the second optical wavelength source may be between 55 and 70 and preferably is substantially 65 . The first optical wavelength source may have a wavelength of substantially 700 nm and the second optical wavelength source may have a wavelength of substantially 575 nm.

According to a second aspect of the present invention there is provided a method of measuring an analyte concentration in a fluid sample, the method comprising the steps of; (i) applying the fluid sample to an area of a substrate comprising one or more reagents that change colour in accordance with analyte concentration; (ii) measuring the reflectance of the area of the substrate at a first optical wavelength; (iii) measuring the reflectance of the area of the substrate at a second - 5 - optical wavelength; and (iv) determining the analyte concentration from the reflectance value at the first optical wavelength and the reflectance value at the second optical wavelength.

Preferably step (iv) comprises calculating a ratio of the reflectance value at the first optical wavelength to the reflectance value at the second optical wavelength and determining the analyte concentration in accordance with the ratio. Step (iv) may further comprise the step of determining the analyte concentration by comparing the ratio with the contents of a database. The method may additionally comprise the additional step of (v) outputting an indication of the analyte concentration to a user.

The invention and its method of operation will be described by way of illustration only and with respect to the accompanying drawings, in which Figure 1 shows a schematic depiction of a device according to the present invention; Figure 2 shows a graphical depiction of the variation of local linearity with time for a conventional single wavelength system at multiple glucose concentrations; Figure 3 shows a graphical depiction of a conventional plot of reflectance with time; Figure 4 shows a graphical depiction of the relationship between the a parameter and the b parameter; Figure 5 shows a graphical depiction comparing the - 6 performance of a test rig according to the present invention with the performance of a test rig using a single wavelength; Figure 6 shows a graphical depiction illustrating the effect of the rotational offset of the optical detector from the specular reflection beam on the performance of a test rig according to the present invention and a test rig using a single wavelength; Figure 7 shows a graphical depiction of the comparison of CV values between a variable single wavelength and a combination of 870nm and a variable single wavelength; and Figure 8 shows a schematic depiction of a second embodiment of the present invention.

Figure 1 shows a schematic depiction of an device 100 according to the present invention. The device 100 comprises receiving means 20, optical source 30, optical detector 40 and control means 50. The receiving means 20 comprises an opening 22 that is positioned and aligned so as to be illuminated by the light that is generated from the optical source 30. Preferably the optical source 30 comprises two light emitting diodes 31, 32 that each emit light at one of the desired optical wavelengths. Both the optical source 30 and the optical detector 40 are in communication with the control means 50. The optical detector 40 is provided with an aperture 42 that reduces the amount of spurious off-axis light that might otherwise be detected by the optical detector, causing inaccuracies in the measurements being made by the device.

Alternatively, a lightguide may be positioned such that 7 - one end of the lightguide is aligned with light reflected from the carrier and the other end is aligned with the detector.

When operating a meter according to the present invention in the indosage application, the user inserts the strip into the device, causing the meter to turn on. Once the optical source illuminates the read zone then the control means 50 will prompt the optical detector to begin to monitor dynamic changes in the ratio of reflectances at the two wavelengths.

An efficient method by which these dynamic changes can be monitored is referred to as a local linearity (LL) check.

A local linearity check requires that the current reflectance (or reflectance ratio) value is taken and processed with two preceding values in order to determine the trend of the values. If rt is the reading at a time t, then rt1 and rt2 are the two immediately proceeding readings and the local linearity can be defined as LL= r, +r'-2 -2r, *100 r, If the LL value is within a given range then this indicates that the strip has been inserted correctly. The LL value can be used as a continuous signal filter to detect operator errors, such as moving the strip during the test. It can also be used to initiate a timing sequence. The purpose of the timing sequence is to define a maximum time for the test, usually less than 20 seconds.

The actual time of the test will depend upon the glucose concentration. The typical shape of the LL signal, which is shown in Figure 2, contains a sharp negative to positive change after sample application (Figure 3 shows a conventional plot of reflectance against time for comparison purposes). The sharp negative to positive change in Figure 2 indicates that the test fluid has reached the area where the chemical changes occur (often referred to as the read zone). The reflectance is continuously measured at the two different wavelengths.

If the light source comprises two LEDs, then it may not be possible to measure the reflectance values at both wavelengths at the same time, and thus the reflectance values can be calculated from a sequence of measurements (this assumes that when the reaction is complete any change in colour will be very slight and can be calculated using a linear approximation).

For the off-meter dosage application the user applies a fluid sample to the carrier outside the meter. The strip is then inserted into the meter and the meter turned on.

In order to make an accurate measurement it is necessary to determine whether the fluid sample has been applied to the carrier for a sufficient period of time in order for the colour change to be complete. This can be determined using the local linearity (LL) technique. When a stable reflectance ratio, or a predicted reflectance ratio, has been obtained the measurement data can be further processed. The reference used is the reflectance ratio between the two analytical wavelengths. This ratio is independent of the initial intensity of the light source - 9 - within the range of glucose concentrations encountered clinically. The reference used is the reflectance ratio when the reaction is complete, and thus it is not necessary to know the original read area reflectance.

As a an alternative to the methods described above he control means may cause the carrier to be illuminated for a set period of time, or until the measurement data received by the control means meets one or more pre determined criteria, such as, for example, passing a threshold value, maintaining a threshold value for a given period of time, two signals attaining a ratio, etc. The control means 50 analyses the received measurement data, preferably by calculating the ratio of the reflectances for the two optical wavelengths and then determining the concentration of the target chemical constituent, either by direct calculation or by reference to a look-up table.

If a ratio of reflectance values is to be calculated then it may be calculated using slightly modified reflection values that are calculated based upon the electrical amplifier's bias (which is estimated at meter production time.

Conventional meters that detect colour changes assume that the optimum wavelength or wavelengths to be used depend on the colour of the signal producing dye and are not affected by the arrangement of the various optical elements.

If the amount of reflected light in an optical arrangement is denoted by 10 reflection= A*a +B*p [I] where A and B are geometry associated constants, is associated with the concentration to be measured (generally linked to the colour of the surface) and a is associated with the rest of the reflected light (generally describes the roughness and reflectivity of the surface).

Conventionally, a relative measurement is used, that is a ratio between the analytical signal and a reference signal is determined. A reason for doing this is to take account of the intensity of the light being used. If the reflection from final coloured surface is compared with the reflection from the surface once the colour has changed, then the geometry related constants A and B will be the same, however a and will be associated with the developed colour stage. The reflectance ratio, r, used to predict the concentration can therefore be described as: A*a+B*,8 A/B a+l5] C*a+3 [2] A*a+B*13 yB*a+,8 C*a+p where the geometric parameter becomes C (with B=O otherwise there would be no way to predict concentration).

Manufactured meters will have slightly different optical arrangements due to mechanical tolerances. Let rO denote the reflection values produced by a reference meter (and subsequent production meters should have similar values to this) and r denote the reflection produced by a subsequent production meter. It is possible to identify the meter differences with the difference between the C constant, which describes the geometrical arrangement, with C - 11 describing the geometrical arrangement of the production meter and CO describing the reference meter. This can be defined by the following equations: C*a+[ ; r0= C - [3] Let f:r-,rO denote the function which describes the relationship between them. As it is intend to use the meters to measure colour variations it is valid to assume; r0()= f(r()) [4] Assuming fis linear (i.e.f(r)=a*r+b), then r0()= f(r())=a*r()+b [5] and thus C0*a+[ =a*C a +[ +b [6] CO *oc+p C*+p Differentiate both sides by and re-arrange to give C*a+p [7] CO *or + which can be applied to [6] to derive b=(C -C)* a [8] If both a(C) and b(C) are linear in C then b(a) will be 1lnear ln a, glvlng C a*(CO*a+)-[ [9] Liz and b(a) = C * (CO a (CO a + /3) - /3) [ 10] which gives b(a)=-*(l-a) [11] a where the gradient of the linear function is: b (a)=-[12] a Figure 4 shows the relationship between the values of the a parameter and the b parameter. The data points were derived from measurements taken with a reference meter and with meters having different geometries. The linearity shown in Figure 5 indicates clearly that the geometric parameter, C, plays an important part in understanding the relationship between the parts of the reflected light and the colour of the surface being measured and that the optical geometry of a meter is very significant in the operation of the meter.

The next factor to be considered in relation to the performance of a meter is the wavelength that is used within the meter. Generally a meter operates by determining a relationship between the concentration(s) to be measured and the received signal(s), (normally a reflectance ratio is the received signal). This relationship is often referred to as the calibration curve(and is often described using the Kubela/Munk transformation, K/S, to linearize the relationship). Let the mathematical representation of the relationship be denoted by T:r-+cor T(r)=c, where r denotes the reflection ratio and c denotes the concentration of constituent to be measured.

In order to determine the precision of concentration - 13 measurements the Coefficient of Variation (CV) values of concentration (cCV) will be compared with the CV values calculated from the underlying reflected signal (rev).

Coefficient of Variation is a relative representation of the standard deviation and is the most suitable representation of the precision of measurements. The cCV is obviously affected by the mathematical transformations that were used to calculate it. When compared to the underlying rev it depends only on this transformation, T. and nothing else. Mathematically this can be shown by: cCV = r |T(r)| cCV r |drT(r)| [13] rCV T(r) rCV T(r) The T transformation, which is dependent upon the design of the meter, will determine the performance of the meter.

For optimum meter performance the meter will provide accurate repeatable measurements (that is, the ratio of rCV to cCV is constant) across the range of concentration values that are of interest. The performance characterization used to estimate the calibration curve is thus very important and may be described by T(rO)=g*rOh [14] where g and h are constants and the O index refers to a reference instrument that generally satisfies that condition that cCV/rCV=constant. Another meter's transformation function (aegis linear) can be written as T(r)=g*(a*r+b)h [15] This causes this meter's performance to vary through the concentration range and the effect can be quantitatised by applying [13] to give - 14 cCV r*h*g*(a*r+b)hl *a r*h*a h [16] rCV g*(a*r+b)h a*r+b 1+ b a*r This constraint can be used to select the wavelength(s) incorporated into the meter design once the geometrical arrangement of the optics has been designed.

For both on-meter and off-meter applications the arrangement of the optical elements is preferably on one side of an axis. As there is a need to reduce the effect of /a, experimental studies suggest that as the angle between the main reflected beam and the optical detector increases, the effect of /a decreases. Conventionally, either the light source or the detector are substantially perpendicular to the read surface, which limits the angle range between the main reflected beam and the optical deflector to less than 90. In practice, it is difficult to achieve angular offsets approaching 90. To achieve larger angular offsets it is necessary to arrange all the elements on the same side of the meter. Using this method it is possible to achieve an angular offset of 180 and in practice it is simple to achieve an offset of 90.

The device then provides an indication to the user of the concentration of the target constituent chemical in the biological material via visual indicators 60 and/or audible indicators 62, for example a numerical indication of the target constituent chemical concentration.

Informed users may then use the data provided by the diagnostic device to manage their condition through diet, exercise or medication, or seek the advice of a medical professional.

In an alternative the optical source 30 may comprise a white light source, or another wideband source that comprises the optical wavelengths of interest, as opposed to a discrete optical source for each wavelength of interest. Similarly, the optical detector 40 may be a wide band detector from which the wavelengths of interest can be extracted, or a discrete optical detector can be provided of the wavelengths of interest, with each detector having a band pass filter centred upon the respective wavelength of interest.

One particular application of a device according to the present invention is the measurement of glucose in blood, as diabetes is becoming an increasing problem as Western populations age and become more prosperous. Theoretical analysis and experimentation have shown that when used with standard reagents for measuring blood-glucose levels it is possible to measure glucose concentrations with a greater repeatability than is possible with conventional measuring devices.

Known measuring devices tend to use a 635 nm optical source to determine glucose concentrations by measuring changes in reflectance at that wavelength. Measuring devices that utilise a further optical source, for example 700 nm, use the second wavelength to correct for other factors, for example haemocrit and oxygenation behaviour within the diagnostic reagents. In the present invention both wavelengths are used to generate reflectance values and the glucose concentration is derived using a ratio of the two reflectance values.

Figure 5 is a graphical comparison between the performance of a conventional diagnostic device using a single wavelength (indicated by triangles in the graph) and the performance of a device according to the present invention using one fixed wavelength and one variable wavelength.

(indicated by squares in the graph). The vertical axis is a coefficient of variation (CV) value, which is a relative representation of measurement standard deviation; the lower the CV value, the more repeatable the measurement technique. The data in Figure 5 for the conventional diagnostic device was generated by taking a number of measurements from 300 nm to 100 nm and calculating a CV value based on those measurements. In each case the reflectance measured was compared to the initial presample reflectance. For the device according to the present invention measurements were again taken across the 300 nm to 1100 nm range. The CV values for each wavelength were calculated using the ratio of reflectance measured at each wavelength with the reflectance measured at 576 nm (the singularity that would be present at 576 nm has been removed from the graph).

Figure 5 shows that the CV values calculated for the device according to the present invention are lower than those calculated for the conventional device across the entire wavelength range, indicating that the device according to the present invention provides more repeatable measurements than conventional devices. An ideal meter would have low and constant CV values across the concentration range that is of interest.

The calculated CV values and hence the performance of the meter has been observed to be dependent upon both the wavelengths used and the angle that there is between the optical detector and the optical source. Figure 6 shows the variation of CV values with the offset between the optical source and the optical for a conventional device measuring reflectance at 700 nm (depicted by triangles) and a device according to the present invention measuring reflectance at both 570 nm and 700 nm (depicted by squares). Figure 6 shows that the optimum offset angle appears to be between 55 and at least 70 and also that the CV values for the device according to the present invention are lower than the CV values for the conventional device. Experimentation has shown that an offset angle of approximately 65 is of significant benefit. Figure 7 shows the variation in CV values for alternative chromagen systems with time.

It will be readily understood that if different diagnostic reagents were to be used to detect glucose concentrations then different wavelengths may be needed so that the device operated in a regional of minimal CV values (that is, the most repeatable measurement of glucose concentration. In the example discussed above a combination of 576 nm & 700nm was used with the DAOS product, made by Dojindo Ltd of Kumamoto, Japan (which comprises 4-Aminoantipiryne and N-Ethyl-N-(2-hydroxy-3 - 18 sulfopropyl)3,5-dimethoxyaniline). If the reagent set used in the QuickTek meter-system (which is 4- Aminoantipiryne and N-Ethyl-N-(2-hydroxy-3-sulfopropyl)- 3,5-dimethylaniline), a product of Hypoguard Limited, Woodbridge, Suffolk, UK, is used then experimentation has shown that the best pair of wavelengths is 570 nm and 870 nm (Figure 7 shows the comparison of CV values between a variable single wavelength and a combination of 870 nm and a variable single wavelength). If the reagent is the o Toluidine reagent (available from Sigma-Aldrich Corp., St. Louis, MO, USA, product no. 6356), then the best pair of wavelengths is 660 nm & 740 nm. The two wavelength values for a particular reagent system can be determined by performing a number of scans using a single wavelength and a range of wavelengths, changing the single wavelength for each different scan. This creates a matrix of CV values for different wavelength combinations and an appropriate choice may be made from within the matrix.

Experimentation has shown that the conventional arrangement of the optical elements is not the most effective. Figure 8 shows a schematic depiction of a second embodiment of the present invention. The device comprises receiving means 220, optical source 230, optical detector 240 and control means 250. The receiving means 220 comprises an opening 222 that is illuminated by the optical source 230 that preferably comprises two light emitting diodes 231, 232 that each emit light at different optical wavelengths. In an alternative embodiment, the light source 230 may comprise a single LED (or other monochromatic light source).

The LEDs 231, 232 and the optical detector 240 are substantially parallel to each other, such that the two LEDs both illuminate the opening 222 in the receiving means and that the detector 240 is aligned with the opening so as to receive light reflected from the sample that is received within the opening. Both the optical source 230 and the optical detector 240 are in communication with the control means 250. It will be understood that the positioning of the LEDs relative to the detector may be varied, for example by positioning both LEDs to one side of the detector.

The second embodiment relies upon the insight of the inventor that the light reflected from the sample is dependent upon two factors. The first factor is the colour of the sample, which is in turn dependent upon the analyte concentration in the fluid. Light will be absorbed by the dye materials within the fluid from the light source 230 and then emitted. The light emitted in this fashion is indicated schematically by distribution 280. The second factor is reflections that are dependent upon Snell's law and the geometry and roughness of the surface of the carrier comprising the sample. The light reflected in this fashion is indicated schematically by distribution 270.

It has been determined experimentally that the light that is reflected due to surface effects is generally of a larger magnitude than the light that is absorbed and emitted by the dye materials and thus conventional meters (that is those using a similar optical geometry to the arrangement shown in Figure 1) are positioned so as to measure the largest signal i.e. the combination of reflected light and emitted light. However, the emitted light is the only light that is of interest and the reflected light is effectively noise. Therefore, the embodiment shown in Figure 8 enables the emitted light to be detected whilst reducing the deleterious effects of the reflected light.

Similarly, the use of the invention to detect other analyses will require the use of different wavelengths to enable the repeatable operation of the device. References to optical wavelengths should be understood to include the visible spectrum as well as the ultra-violet and infra-red spectral regions near to the visible spectrum. Although a meter according to the present invention may be a portable device for use by non-medical personnel, the present invention may also be provided as a larger device, for example a bench-top machine, that is intended for use by medical personnel, for example within a doctor's surgery or a hospital. - 21

Claims

1. An apparatus for measuring analyte concentration in a fluid sample, the apparatus comprising: a first optical wavelength source and a second optical wavelength source; an optical detector and an opening, the first optical wavelength source and the second optical wavelength source being aligned, in use, to illuminate a fluid sample received within the opening, the optical detector being aligned, in use, to measure light reflected from the fluid sample at the first optical wavelength and the second optical wavelength, the control means determining the analyte concentration in accordance with the reflected light measured at the first optical wavelength and the reflected light measured at the second optical wavelength.

2. An apparatus according to claim 1 wherein the first optical wavelength source, the second optical wavelength source and the optical detector are all substantially parallel.

3. An apparatus according to claim 1 wherein the first optical wavelength source and the second optical wavelength source are substantially parallel and there is a rotational offset between the optical detector and the first optical wavelength source and the second optical wavelength source.

4. An apparatus according to claim 3 wherein the rotational offset between the optical detector and the first optical wavelength source and the second optical - 22 wavelength source is between 55 and 70 .

5. An apparatus according to claim 4 wherein the rotational offset between the optical detector and the first optical wavelength source and the second optical wavelength source is substantially 65 .

6. An apparatus according to any preceding claim in which the first optical wavelength source has a wavelength of substantially 700 nm and the second optical wavelength source has a wavelength of substantially 575 nm.

7. An apparatus for measuring analyte concentration in a fluid sample substantially as described hereinbefore and with reference to Figure 1.

8. An apparatus for measuring analyte concentration in a fluid sample substantially as described hereinbefore and with reference to Figure 6.

9. A method of measuring an analyte concentration in a fluid sample, the method comprising the steps of; (i) applying the fluid sample to an area of a substrate comprising one or more reagents that change colour in accordance with analyte concentration; (ii) measuring the reflectance of the area of the substrate at a first optical wavelength; (iii) measuring the reflectance of the area of the substrate at a second optical wavelength; and (iv) determining the analyte concentration from the reflectance value at the first optical wavelength and the reflectance value at the second optical wavelength.

10. A method according to claim 9, wherein step (iv) comprises calculating a ratio of the reflectance value at the first optical wavelength to the reflectance value at the second optical wavelength and determining the analyte concentration in accordance with the ratio.

11. A method according to claim 10, wherein step (iv) further comprises the step of determining the analyte concentration by comparing the ratio with the contents of a database.

12. A method according to any of claims 9 to 11, the method comprising the additional step of (v) outputting an indication of the analyte concentration to a user.