CA2376747A1

CA2376747A1 - Apparatus for quantitative in vivo optical measurement of a mammalian analyte concentration and use thereof

Info

Publication number: CA2376747A1
Application number: CA 2376747
Authority: CA
Inventors: Stig Ollmar; Peter Aberg; Parisa Baniasad; Per Hellstrom
Original assignee: Individual
Current assignee: Individual
Priority date: 2002-03-13
Filing date: 2002-03-13
Publication date: 2003-09-13

Abstract

An instrument in which NIR light reflects against the skin and the water content correlates with the amount of light absorbed. Water dominates large blocks of the NIR absorption spectra. 1940 nm and 1460 nm were used as measurement wavelengths. A quartz-halogen lamp was used as a light source for the instrument and two band pass filters were used to select the two measurement wavelengths. Laser diodes can also be used. With laser diodes the wavelength is fixed and the light can be chopped by switching the laser diode on and off rapidly. Chopping is a convenient way to synchronise he photo diode (the sensor) with the source. Tests on thin layers of paper show the feasibility of the approach described; for a relatively inexpensive, non-invasive instrument for water content measurement in skin.

Description

8 a- , ;fir ,_r.. ~, r APPARATUS FOR QUANTITATIVE IN VIVO OPTICAL MEASUREMENT OF A
MAMMALIAN ANALYTE CONCENTRATION AND USE THEREOF
FIELD OF THE INVENTION
This invention relates to an apparatus for quantitative in vivo measurement of a mammalian analyte concentration, an example of the analyte being water (moisture) within skin of a human subject. The apparatus is a spectrometer which operates in the range of infrared light. The invention also relates to uses of an apparatus, as in a method of quantitatively determining the amount of analyte and further, to the use of information obtained in such method in the aid of making further determinations with respect to the subject, such a determining blood glucose level by measux-ing skin impedance.
BACKGROUND OF THE INVENTION
NIR Spectrophotometry In the past few decades, there has been an increasing interest in applying Near Infrared Spectrophotometry (NIRS) to the problem of food and beverage analysis [Osborne 1993].
NIRS is the use of near infrared spectra of electromagnetic waves for the analysis of the structural bonds in the molecules of a sample material. NIRS has a broad range of applications from maturity evaluation of fruits to automatic selection of defective products in a production line. For example, NIRS techniques have successfully been employed in biomedical, pharmaceutical, agriculture, chemical, polymer and petroleum industries [Kumar 1996], [Osborne 1993], to name just a few.
NIR investigations include various applications ranging from the measurement of haemoglobin [Kumar 1996] and glucose [Burmeister 1998] in blood to the determination of octane number in gasoline samples. The NIR technique is well suited for process control testing and particularly for on-line t- ~_ r .

applications because analysis can be done rapidly and non-destructively [Bokobza.l998].
The NIR spectrum, which covers the range between approximately 700 nm (near the red end of the visible spectrum) and 3000 nm (near the beginning of infrared stretches of organic compounds), contains overtones and combination bands that are derived from the fundamental absorption known from infrared (IR) spectroscopy. Combination bands arise from combinations of fundamental C-H, N-H and O-H
vibrations [Dempsey 1996] .
The number of combinations and overtones in an NIR
spectrum are large, resulting in an overlapping series of bands that is rather smooth. Bonds that vibrate at high energy and with large amplitude carry most intensity in particular those involving hydrogen, the lightest element.
The overtones and combination bonds are much weaker (usually by a factor of 10 to 100) than the fundarr~ental absorption bands. Therefore, analysis of samples that are several millimetres thick is possible [Bokobza 1998].
A potential applications of NIR is in measuring the moisture level of biological samples. The application of NIR
for moisture measurement is not new. Norris investigated its use for grain analysis [Osborne 1993].
Absorption of NIR radiation by the sample changes as a function of wavelength, and this phenomenon makes it possible to study different parts of the NIR spectrum for different purposes. This means that it is possible to measure content of carbohydrates, proteins and water in samples by the study of the NIR spectrum. During the measurement the chemical information of he sample repeats and the overtones of fundamental absorption give the most speci~ral information.
Since NIR spectrophotometry is non-destructive and can be performed in an on-line fashion, it is a particularly ideal s , .-~t w: l r candidate for measuring the moisture level of biological samples. Modern near infrared spectrometers have improved to the point where it is now possible to analyse living human tissue in vivo for the clinical diagnosis.
NIRS has many advantages and among them the simplicity of sample preparation for liquids, solids or gases should be mentioned. In clinical and chemical measurements, reagents are needed and through the recording of the sample's spectrum NIRS allows a multicomponent analysis of many samples at the same time [Heise 1999]. In comparison to many of the more conventional analytical methods analysis, analysis time can be reduced to less than one second and in this way this technique saves time and materials. It also facilitates a non-invasive and non-destructive analysis, and estimation of physical properties and biological effects from the spectra of samples.
Sample sizes have a very big range from picograms to planets, and in each case molecular structural information can be derived from the spectra. The method is relatively inexpensive [Dempsey 1996]. A disadvantage of NIR is the weak sensitivity to minor constituents [Bokobza 1998].
BRIEF DESCRIPTION OF THE DRAT~JINGS
The present invention is described with reference to the following figures:
Figure 2.1 is a schematic illustration of the normal structure of the human skin [Martin 1999].
Figure 3.1 shows interaction of radiation with the matter [Osborne 1993 ] .
Figure 3.2' is a graphical statement of the Beer-Lambert law. A = abc showing the effect of deviai~ion due to polychromatic radiation, chemical interactions (1) and scatter, reflection (2) [Osborne 1993].

r , J r ~ X r Figure 3.3 shows tangent (1) and baseline (2) methods for background correction [Osborne 1993].
Figure 3.4 shows the spectrum of a mixture of two analytes with overlapping peaks (7~1 and ~,2) [Osborne 1993] .
Figure 3.5 shows resolution of two bands and the convolution from a linear background by second order differentiation (dashed line) [Osborne 1993].
Figure 3.6 shows reflection from smooth and rough surfaces [Osborne 1993].
Figure 3.7 illustrates diffuse (body) reflectance [Osborne 1993].
Figure 3.8 shows an NIR spectra of three flour samples with the same composition but ground to three different particle sizes [Osborne 1993]
Figure 3.9 shows an NIR spectra of a flour sample, before (1) and after (2) oven drying [Osborne 1993] .
Figure 4.1 shows a double FP interference filter.
Figure 5.1 illustrates optical pathways for light in skin: (a) absorbance or transmission, (b) regular reflectance,, (c) diffuse reflectance, and (d) multiple internal scattering [Martin 1999] .
Figure 5.2 shows the basic instrument configuration for reflectance measurements used herein.
Figure 5.3 shows that absorption by melanin is negligible above ,110 0 nm . [Anderson 19 81 ] .
Figure 5.4 shows a chopper wheel of a spectrometer of the present invention, (1) - (3) Synchronise holes, (4) - (5) filters and ( 6 ) hole .
Figure 6.1 shows the output of the band reject (notch) filter is ebr [Clayton 1979] .
Figure 7.1 is a bar graph showing output voltages for dry and wet paper towels. Measurement wavelength is 1940 nm and the T-signs over the bars represents the noise. There is a significant difference between the dry and wet states of the paper towels.
Figure 8.1 is a schematic of an instrument amplifier with matched photo diodes.
Figure 9.1 illustrates the LabView programme:
DETAILED DESCRIPTION
Skin is the largest organ in the body. Its primary function is as a protective barrier between the body and the external environment. This multiple layer organ with a surface area about 1.8 m2 consists of two major parts: the epidermis and the dermis. A linking complex structure, called basal lamina or basement membrane, connects the epidermis and the dermis (Figure 2.1).
The first part, the epidermis, is a thin membrane with approximately 100 ~m thickness and has no vessels. In contrast to the epidermis, dermis has vessels and is composed of connective tissue. Epidermis acts as a barrier. This layer blocks entering of free water and foreign substances into the body and at the same time prevents excessive water loss from the body. Various physical and chemical environmental factors affect the most superficial part of the epidermis, the stratum corneum. These give rise to a continuous destruction and desquamation, partial loss of the epidermal basal cells. To compensate for this, a continuous renewal of the epidermis takes place. The turnover time for the replacement of the lost sub-layers is thought to vary between 52 to 75 days [Nicander 1998].
As described by Nicander, cells in the epidermis form four layers or strata: "basal cells (stratum basale), prickle cells (stratum spinosum), granular cells (stratum granulosum), and keratin (stratum corneum or keratinized layer)".

7 f T i a <

Stratum corneum (SC} is the outermost layer of skin and like the other parts of epidermis consists of keratinocytes.
The keratinocytes, which are the largest cells of the epidermis, lose all their organelles under their lifetime.
They even lose their nuclei and water. These dehydrate cells have a shape of a hexagonal disc and are very flattened [Nicander 1998]. These cells are filled with a kind of insoluble protein called keratin, which is a tough substance.
There is even a large amount of keratin in nails, hooves, horns and feathers. The space between the cells contains smaller amounts of lipids and water [Martin 1999].
As mentioned above, a function of the skin is to protect the body from chemical and physical agents, SC has a role in defending the body against not only bacteria and other dangerous organisms and foreign substances but also against ultra violet (UV) radiation. SC's thickness varies from 10 ~m at one place to 40 ~m at another place arid on average is about 15 ~m [Martin 1999], but may increase up to several hundred ~m e.g. the bottom of the foot.
One reason the level of moisture in SC is of interest is to monitor alterations during development of skin diseases, such as contact dermatitis. Another reason is to investigate the effect of creams, lotions, soap and other personal hygiene products on this layer. Dry skin, low water content in skin, is a cosmetic concern. SC's moisture depends on several factors such as the ambient humidity, exposure to W
radiation, exposure to detergents and the general health of the tissue. With the knowledge of skin moisture, it is possible to find suitable products or treatments for the individuals.

T,, ,~ ~"
-Measuring the level of skin moisture There are several methods for studying water in skin in vivo. These methods include electrical conductance, electrical capacitance, transepidermal water loss (TEWL), viscoelastic measurements and infrared spectroscopy by attenuated total reflectance (ATR) [Martin 1998]. According to Martin, all these methods have some disadvantages as described below.
Conductance measurements are influenced by many factors that can introduce uncertainties into the measurements. Salt content, degree of contact with the probe or water content is some of them. Relationship between water in skin and evaporation of water from skin is very complex. TEWL
measurements measure how fast skin loses the water, and viscoelastic methods have an indirect and non-linear relationship to water content. As. mentioned above, skin's moisture depends on many factors and all these cause complexity.
In ATR, a sample is placed on an infrared transparent crystal of high refractive index. The degree of contact between skin and reflectance crystal affects the measurement, as do refractive index. These will vary if skin surface or ambient condition changes. In addition, the depth of penetration of radiation is quite shallow and is restricted to the uppermost SC.
Martin et a1. and de Rigal et a1. have done some studies in the use of near-infrared (NIR) reflection spectroscopy to determine skin's moisture. Earlier studies by de Rigal et al.
were in vitro and later studies were in vivo. During in vitro studies fibre optics were used to show relationship between the water peak intensities and water content. Their in vivo studies indicated there to be a strong correlation between the dryness and water content of skin. Martin et a1. studied NIR

D T. S i n , water bands and the observed scattering effects. They also compared water content at different anatomical sites [Martin 1998] .
Nicander and Ollmar presented a new measurement system based on electrical impedance. This method is an objective and non-invasive method for assessing alterations in the human skin and oral mucosa, which reflects structural and chemical changes in living tissues [Nicander 1998].
Osborne has discussed the physics of the interaction of radiation with matter [Osborne 1993]. He writes, "when monochromatic radiation interacts with a sample it may be absorbed, transmitted or reflected ". A comprehensive discussion on NIR instrumentation is provided by Osborne [Osborne 1993], some details provided herein. In accordance with the law of conservation of energy, the total radiant power incident on the sample (Pp) must be equal to the sum of the radiant power absorbed (PA), transmitted (PT), and reflected (PR) .
Pa _ pA + PT + PR ( 3 .1 ) The significance of this relationship in spectrophotometry is that the radiant power absorbed may be deduced by measuring either the reflected or transmitted radiant power when the other one is zero in ideal conditions under which the experiment is performed.
Transmission of radiation Interaction between the electric vector of the incident radiation and the sample particles cause the electrons contained in the sample particles to oscillate respect to their nuclei, and this leads to a temporary polarisation.
Absorption of radiation changes the frequency of the incident ,f ~ ,' radiation while just a polarised sample :leaves the frequency unaltered. Regardless, when radiation travels through different media the velocity of the radi<~tion is altered. The ratio of the velocity in vacuum to that :in a medium is called the refractive index (n) of the medium. When the radiation propagated in a less dense medium with refractive index n1 crosses the boundary of another dense medium with refractive index nz, an abrupt change in the direction of propagation occurs (Figure 3.1) and this is known as refraction. Snell's law defines this departure sin81 _ nz sin9z n1 (3.2) Where A1 and 82 are the angles of the beam from the normal in each medium.
Absorption of radiation The radiant energy propagated through a sample can be absorbed by various substances in the sample. The amount of absorption of radiant energy at each frequency depends on the kind and amount of the substance in the sample. In other words, each substance has the capacity to absorb the radiation at specific wavelengths. According to the Beer-Lambert,law, which describes attenuation of the transmitted radiation by an absorbing sample, the amount of radiant energy absorbed by a very thin layer of the sample is proportional to the number of molecules in this layer.
- dP/P = k do (3.3) where P is radiant energy and do is the number of molecules in the layer.

n,_ re For measuring absorption through the whole sample, which lies in the path of the radiation, equation 3.3 is integrated;
yielding -In (PT/Po) - kn (3.4) where Po is the power of the incident and PT that' of transmitted radiation, and n is the number of molecules in the path of the beam.
To remove the negative sign in equation (3.4), the ratio of PT to Po could be inverted. Since the.number of molecules in a definite thickness is proportional to the concentration of molecules, n in the above equation can be replaced by thickness, b, multiplied by concentration, c, when a new constant, a, replaces k.
log (Po/ PT) - abc (3.5) The logarithm is now to base 10 and 'c' is expressed in mol-1, b in cm and the constant 'a', which is called the absorptivity, 8 in mol-lcm-1. The logarithm of converted transmittance, which is the fraction of radiation (P/Po) transmitted by the sample, is absorbance A.
A - log (1/T) - log (Po/ PT) (3.6) And equation (3.5) becomes A = abc (3.7) While the Beer-Lambert law holds, the absorbance and concentration have a linear relationship where ab is the slope of the straight line through the origin in the plot. Since the sample thickness is accurately defined, the value of absorptivity may be determined easily by using a series of samples of known concentration. However, despite linearity m si ' s~

between absorbance and path length at a fixed concentration of absorber, deviations are frequently encountered (Figure 3.2).
The deviations from the Beer-Lambert law are the result of ignoring some factors in the beginning. In equation (3.1) reflection is neglected, i.e. PR = 0, but if it is significant it is not sufficient to measure just PT to deduce PA. Another neglected factor is the scattering which increases absorbance A by increasing the effective path length. It is also significant to hold the slit width constant, because decreasing slit width improves the resolution of absorbance.
Other reasons for deviation are chemical interactions such as association, dissociation or reaction of the absorbing species with the solvent or other solutes. Electrostatic attraction between the molecules changes the absorptivity and this is a limitation for a sample with concentration above about 0.01 mol: To derive equation (3.7) monochromatic radiation is assumed which affect the relation in two ways. It is assumed that there is a dependency between the absorptivity and wavelength. For a given wavelength, absorptivity has a maximum peak but for a broad peak the absorbance variation will not be seriously dependent on wavelength. It is also assumed that there is interference between the monochromatic radiation and impure radiation reflected from various surfaces in the instrument without going through the sample. In this condition the observed absorbance is given by A~ log P +P (3.8) T s where PS is the power of the stray radiation.
The Beer-Lambert law forms the basis for quantitative absorption spectroscopy and a given system would not adhere to it because of various interferences in the measurements. The line in Figure 3.2 is assumed to go through the origin, but a a ~ a a ' x this assumption is not always correct. 'The relationship between the concentration and the absorbance can be represented by A = Ao + abc ( 3 . 9 ) The Beer-Lambert law does not hold for analysis of food slurries by NIR transmission spectroscopy for two reasons: the length b is no longer a constant because of the effect of scattering, and log (1/T) does not represent the attenuation of the beam by absorption, since some of the radiation undergoes diffuse reflectance.
The reading must be corrected for background absorbance to get a true reading for the desired,analyte. The background correction can be achieved in different ways such as double beam spectrophotometry, the cell-in, cell-out method and the baseline method. As a reference absorber, the solvent is used in the first two methads either in a matched cell in a reference beam, or in the same cell read immediately after the sample. Absorbance, A, is calculated by logarithm of (Psolvent/Psolution) . Baseline correction is more general, and in simple situations the absorbance due to the background interference shifts the baseline. To diminish the effect of the background it is necessary to subtract the reference reading, the minimum absorbance in the spectrum, from the desired peak maximum i . a . corrected absor:bance = A~,max - A7~,n,in -DA. But it is insufficient simply to use just the spectrum of the interference if the absorbance of the interference is not equal at a.,naX and 7~.min. The linear background has to be averaged (Figure 3.3).
For the contaminants with known spectra, however, other wavelength than a.min maybe used. Drawing a tangent across the base of the peak, the tangent method (Figure 3.3), may be utilised for a not constant but linear background. The t a k s tangent method is only valid when the composition changes across the base have no effect on the tangent.
If the peak of the contaminant overlaps with the peak of interest the law of additivity is still valid but the correction technique is more complex and the analysis of a multicomponent system may be involved. Gne material does not affect the absorbance of another according to the law of additivity, The absorbance for each material is calculated by equation (3.7) while the observed absorba.nce at one given wavelength is a linear sum of all the individual absorbances.
A = bE ai Ci (3.10) where ci is the individual concentration and ai is the individual absorptivity. This situation is illustrated in Figure 3.4 when the spectrum is a mixture of two component x and y.
The absorbance of the mixture A1 and AZ at the two wavelengths ~,1 and 7~2, where x respective y has the maximum, may be described by the two equations A1 = alXb cX + alyb cY ( 3 . 11 ) AZ = a2Xb CX + a2yb Cy ( 3 . 12 ) If the Beer-Lambert law is followed by both components these equations may be solved for cX and cY, the concentrations of x and y in the mixture, because the path length b, the four absorptivities alX, aly, a2X; aaY and the absorbances A1 and A2 are all experimentally determinable.
It has been shown that for the NIR spectrophotometric analysis of mixtures of hydrocarbons (Liddel and Kasper 1933) and amides (Krikorian and Mahpour 1973) and for water bound to protein (Hecht and Wood 1965) the Beer-Lambert law is valid.
On the other hand hydrogen bonding in the mixture of methanol and aniline causes deviations from additivity [Osborne 1993].

b , , r T , Using the above equations makes it possible to estimate different components in a mixture, but in a difficult way since for each component an experiment at a given wavelength is needed. For example, for estimating four components in a mixture, four experiments must be done at four different wavelengths fairly well separated.
The problems of overlapping peaks anal baseline correction can be approached in other ways such as using derivative spectra (Giese and French 1955). In Figure 3.4 the absorbance A is shown as a function of wavelength ~,. To calculate the derivative spectrum (dA/d~,) from a tabulation of absorbance at equally spaced wavelengths, the difference between absorbance at adjacent wavelengths is computed. In practice some degree of smoothing is needed. A first derivative spectrum has a peak where the upwards slope in the original spectrum reached the maximum, a trough where the downwards slope of the original reached a maximum, and a value of zero at any peak in the original.
The second derivative spectrum (d2A/d~,2), which is more interesting, is obtained by differentiating the first derivative spectrum with respect to ~,. Where the points have maximum curvature in the original spectrum there are peaks and troughs in the second derivative spectrum, in other words there is a.trough for each peak in the original. Figure 3.5 shows the spectrum of Figure 3.4 together with its second derivative. The linear background is converted to a constant level in the first derivative (slope) spectrum and zero in the second derivative (curvature) spectrum. In the same way a quadratic background in the original spect=rum would have become linear in the slope and constant in the curvature spectrum. A disadvantage of this however is a deterioration in the signal to noise ratio since the dif-_ferentiation operation amplifies the noise and increases the complexity of the spectra. The third, fourth or higher derivatives are rarely used because the deterioration increases with each derivative. Since differentiation is a linear operation, if the Beer-Lambert law was obeyed in the original spectrum it would imply linearity with concentration in the derivative spectra of all orderse d~ ~ dal," l be ( 3 .13 ) Hence the derivative spectra and the original spectra can be used quantitatively in the same way.
Reflection of radiation The Fresnel equation describes the specular reflectance (also called regular reflectance) that takes place at the sample-air interface. When the incident radiation meets the sample surface at a right angle the fraction of radiant power reflected (PR / Po) can be calculated by Px _ (~z -ni)z ( 3 . 14 ) pp (hz + n~)2 where n1 and n2 are the refractive indices of the two media.
Absorption of incident radiation by the sample affects the refractive index by an absorption constant k as imaginary part of the complex number n(1- ik) where n is the real part.
Px _ (nz -h~)z +(hzk)z p0 (~z +~')z +(~zk)2 (3 . 15) If a sample is non-absorbing and opaque this equation no longer applies, instead the Laws of mirrors describe the reflected beam of the incident radiation. According to these laws, both the incident beam and the reflected beam as well as the normal to the mirror at the point of incidence lie in the same plane, whereas the angle of incidence is the same as the a <.

angle of reflection. A rough sample may be assumed as a series of small surfaces directed in different orientation at all possible angles to the normal. Fresnel's law is valid for every reflected beam (Figure 3.6).
Sometimes a part of incident radiation undergoes absorption through the first interface while the remaining part transmits to the next interface according to the Beer-Lambert law where it may be reflected, absorbed or transmitted (Figure 3.7) .
If the particles within the sample are much larger than the wavelength, the radiation would be reflected in all directions and create a phenomenon known as scattering, which was discovered by Tyndal in 1869.
Therefore, a sample such as flour, which has a thick layer, causes multiple scattering and produces a diffuse white reflection. Contrary to optical measurement where scattering causes error, scattering in biological samples has intensification effect and gives rise to more absorption of radiation. There are no strict theories for diffuse reflectance due to difficulty of mathematical description of light propagation in the sample, but there are some proposals, among them one by Kubelka and Munk [Kubelka 1948]. They defined the Kubelka-Munk function, F(Roo)~ to describe the power of reflected radiation by scattering, s, and absorption, k, constants in the special case when the sample was opaque and had an infinite thickness. They proposed the following equation:
(1-Roo)2 _ k 2Reo s ( 3 . 16 ) where Roo is the reflectance of the infinitely thick layer.
Some studies have been done to verify this equation (Butler and Norris 1960; Kortum et a1. 1903; Kortum 1969; Law n n Y. l Y f' '. f and Norris 1973) and it has been mentioned that there is a relation between the surrounding medium with the scattering coefficient and the refractive index. Kartum et a1. showed that a 1 cm close packed thick layer is enough to satisfy the condition of equation (3.16). In order t:o relate this equation to sample concentration, all the diffusely reflected radiation must be collected and the absorption constant k be replaced with the absorptivity (defined by the Beer-Lambert law) multiplied by the concentration [Kortum et al. 1963;
Birth and Zachariah 1976] as described by the following equation:
Roo = (1 Roo)2 - _ac ( 3 . 17 ) 2Roo s In food analysis, the Kubelka-Munk function, F(Roo), is defined by the logarithm of the ratio of a non-absorbing standard as a reference to the diffuse reflectance, where the logarithm makes the relation between the concentration and the diffuse reflectance nearly linear (Morris 1983).
log (R'/R) - log 1/R + log R' ~ ac/s (3.18) where R' is the reflectance of the standard and R that of the sample (R' > R).
Equation (3.18) could be rewritten to represent a mathematical model for the use of NIR spectroscopy in the food analysis when monochromatic radiation is used (*) and the law of additivity (**) is valid.
(*) log R' - constant and can be neglected (**) log 1/R - (1/s) ~aici c = k + (s/a) log 1/R (3.19) The "scattering constant", s, is not a constant and varies according to a number of sample properties such as particle size and moisture. A sample with a relatively large particle size lets the radiation penetrate deeper into the sample and diffuses less. This means that s has an inverse relation to the particle size (s oc 1/d) .
Reducing the diffuse reflectance increases log 1/R.
Figure 3.8 shows the diffuse reflectance spectrum along the y-axis as a function of particle size. Figure 3.9 shows that the spectrum is not constant and that~the: scattering power also depends on the wavelength, beside the sample properties.
As mentioned above, water is the another important property of the sample which could affect the spectrum. The scattering, which depends on the ratio of the refractive index of the particle, n, to that of the surrounding medium, no, would be changed if, instead air, water fills the space between the sample particles. Replacing air with water reduces the n/no ratio because of the greater refractive index of water relative to air. Decreasing ratio n/no diminishes the scattering which increase the log 1/R as seen in Figure 3.9.
Since s changes in a complex manner with wavelength for each specific sample, this leads to a difficulty with utilising diffuse reflectance in the analysis. In the NIR
spectroscopy of scattering samples the measurement and the reference wavelengths are chosen carefully to make s nearly equal so that the s/a term in equation (3.19) becomes a constant.
Instrument Components As mentioned above, NIR has a wide range of applications.
When the sample material is an opaque solid, the transmitted constituent of the radiation is zero and the diffusely scattered radiation can be measured. For translucent liquids the reflected portion of the radiation is negligible in comparison with the transmitted portion, and arrangement for the measurement of the latter is desirable.
An NIR instrument is, typically, composed of a source of radiation, detector, chopper and interference filter. A brief description of each component is provided below. A more comprehensive discussion of NIR instrumentation is provided by [Osborne 1993].
Sources In general, the sources of an NIR system may be classified as thermal and non-thermal. Thermal sources typically employ a braad emitting source of radiation such as a tungsten filament bulb. Non-thermal sources use a much narrower beam of radiation similar to those generated by lasers, laser diodes, light-emitting diodes (LED), and discharge lamps.
Thermal In general, NIR spectroscopy requires a broad source of energy. This broad range may span from the visible spectrum to 3 Vim. In the thermal sources, although the distribution of the energy is not critically important, under ideal experimental conditions the distribution should be rectangular and the power output should be stable.
In practice, all NIR instruments employ tungsten filament bulbs as the source of energy with operating power in the range of 10-2OOW.
Despite widespread use, tungsten filament bulbs have disadvantages and limitations', deposition due to evaporation of tungsten, to name just one. To reduce the effect of the deposition of tungsten, instrument designers add an envelope to the filament. This, in turn, presents problems where compact design is an issue.

.. '" it Quartz halogen (QH) bulbs address the problem of deposition of tungsten. In QH bulbs, the evaporated tungsten combines with halogen gas and deposits back on the hot filament. This permits the filament to operate in even higher temperatures with much higher running hours. Normal QH bulbs operate with power requirements in the range between 6V, 12 W
and l2 V, 50 W. Commercially available QH bulbs may run in temperatures within the range of 2600 to 3200 K.
Noa-thermal The sources in this category use a narrow range of wavelengths sometimes as narrow as an individual emission line, reducing the overall power requirements of the instrument. This is achieved by reducing the power consumption of two major parts within the system: the source of the radiation; and the modulation part. Laser and laser diodes require much less power in comparison with their thermal counterparts. NIR instruments that employ non-thermal sources, unlike those with thermal ones, do not require chopping since modulation can be achieved electronically.
This, in turn, eliminates the need for the chopper, a mechanical part with considerable power and size requirements.
Non-thermal NIR systems may be sufficiently power efficient to enable a very compact design compact and one that is battery-powered.
Light Emitting Diodes (LED) Light emitting diodes are commonly used in communication industry. They are typically inexpensive for industrial standard wavelengths. However, for non-standard wavelengths they become much more expensive. LEDs are very low-powered and as a result they produce a weak output. signal. In order to obtain an adequate signal to noise ratio a cluster of LEDs is normally needed.
Laser Diodes The principles of operations of laser diodes are similar to those of the lasers. However, they use wider line-widths and function in the semiconductor scale. Similar to LEDs with interference filter laser diodes' operating wavelengths are fixed. This limits the range of applications that can employ laser diodes and restricts their use.
Lasers Light amplification by the stimulated emission of radiation, or the laser, has various applications in spectroscopy. However, the use of laser for NIR has not gained wide acceptance. The disadvantages of LEDs and laser diodes with regards to dedicated wavelengths and high cost are both inherent in laser: It is expected that a low cost, reliable tuneable laser may become a commercial reality in the relatively near future.
Detector There are two classes of infrared detectors: photon detectors, and thermal detectors. Thermal detectors respond to the whole range of wavelengths and are not efficient enough for practical applications. Photon detectors are more selective and are more preferable for serious use in the NIR.
There is a number of selections for photon detectors including silicon and germanium photodiodes, lead sulphide (PbS), and lead selenide (PbSe). Silicon detectors are sensitive to the visible part of the spectrum and up to 1 Vim.
The peak response of silicon detectors is around 0.85 ~m whilst germanium detectors peak at 1.3 Vim. PbS and PbSe are the most widely used NIR detectors and are used in the range of 1-2.5 ~,m and 2.5-3.5 ~,m, respectively. PbS and PbSe are commercially available in square forms from 1 mm2 to 1 cm2.
They operate in room temperature or cooler, however, in cooler temperatures their response time and signal to noise ratio increase.
Interference filters A multilayer dielectric interference filter consists of a transmitting substrate which supports a series of coatings.
Figure 4.1 is a cross-section through a double Fabry-Perot filter (FP) in which the following elements are represented as follows:
substrate 12 reflecting stacks 14 decoupling layer 16 compensating layer 18 Fabry-Perot cavities The first coating in a double FP filter is a reflecting stack consisting of high and low index materials. The reflectivity is approximately 95 per cent. The first cavity is a low index material 3x ~,/2 wavelengths thick, where ~, is centre wavelength. A second reflecting layer follows this and a ~./4 decoupling layer is next.
The whole process is then repeated for the second FP.
Several cavities may be combined to produce a sharper cut-off and to alter the shape of the passband. 'The last layer, which compensate the decoupling layer, is a. low index 7~/4 coating.
By including low- and high -pass filters (cut-off and cut-on) the design is completed. Absorbing and reflecting layers are also included in the stack to block the transmission of unwanted wavelengths over a wide spectrum from near UV to far IR. There are many other designs of interference filters reflecting both required transmissions characteristics and choice of available materials.
Chopper A chopper is a mechanical rotating disc with slots and holes used in a thermal NIR system for modulation purpose.
The function of a chopper is to intercept light at periodic intervals.
Instrument design Two things may occur when near infrared radiation of wavelengths, which can be absorbed by O-H bands, comes in contact with skin. It can either be absorbed by the water or be reflected. Absorption of NIR radiation causes W bration in individual bonds and the amount of absorbed light depends on the moisture level on skin as described by the Beer-Lambent law and has an inverse relationship with the reflected light.
To visualise the optics of normal skin, it is helpful to schematise the optical pathway through the skin. Scheuplein discussed the optical pathway through skin [Martin 1999]. He defined some possible pathways for light from the W to the infrared through SC. Figure 5.1 illustrates these pathways.
As discussed above; a measuring system combines a source, filters, chopper and detector with sample, signal processing and readout. The arrangement of these components makes it possible to measure transmission, reflection or absorption of the light. Here, light reflected from the skin is measured. A
basic configuration for reflection is out:Lined in Figure 5.2.
The NIR system provided here is divided into essentially four parts: the source, the chopper, the probe and a laptop computer.

9 a ~ ~ 1 D.?

Source The optical properties of stratum corneum have a significant role in absorption of radiation. The transfer of optical radiation depends on the structures and chromophores of SC which varies between different individuals. The thickness, composition and morphology of the SC are some of the modifying factors in absorption or scattering of radiation.
Another significant factor in penetration of radiation in skin is wavelength. Different wavelengths across the optical spectrum, from approximately 250 nm in the ultraviolet to approximately 3000 nm in the infrared, reach different depths within skin. Thus, choice of wavelength can affect the depth of penetration of radiation. Among all wavelengths in near infrared region two wavelengths were chosen: 1940 nm (combination) and 1460 nm (first overtone of the OH stretch) , where the skin has the maximum absorption intensities [Eisenberg 1969]. Another reason in contributing to this selection is that light absorption of the skin pigmentation melanin is small above 1100 nm (Figure 5.3). Over this wavelength the light does not make any difference between white and black skin [Anderson 1981].
There are different sources such as LEDs, laser diodes, lasers and tungsten-halogen lamps that radiates these two wavelengths.
For non-standard wavelengths the light emitting diodes, which can be inexpensive, became much more expensive. To obtain the desired wavelengths these diodes must be combined with interference filters. Another disadvantage of LED for the purposes described herein is the weak output signal of the LED. To improve output signal several LEDs are needed.
Laser diodes produce adequate linewidths for NIR and in contrast to LED their output energy is of the order of several r 1 r ~ T. o hundred milliwatts. Laser diodes have the advantage of wavelength selectivity and the ability to switch the diode on and off with high frequency to obtain a chopping effect.
Optical chopping is useful to improve the signal to noise ratio and is done by detecting the difference between signal and no-signal using a synchronous detection principle. Laser diodes are most common in NIR at the telecom industry wavelengths of 1310 and 1550 nm, require Peltier cooling and laser diode drivers. Another drawback of laser diodes is the effect of temperature on the wavelengths and the lifetime.
Lasers produce a narrow linewidth that is suitable for NIR
measurements but are prohibitively expensive. All of these sources require cooling.
LEDs provide relatively poor wavelength precision and suffer the disadvantage of the need for multiple LEDs to compensate for the lack of radiated power.
Most NIR instruments employ tungsten filament bulbs as a source and this is cheaper than the others sources. The instrument described herein includes a three watt tungsten-halogen lamp manufactured by CVI Laser called ASB-W-003. The lamp is employed as a thermal source to provide the broad emission of electromagnetic radiation from the visible to the near infrared region. It was designed to produce the maximum illumination from a Black Body source into a fibre bundle.
While measuring water in skin the power of the lamp should not be too intense in order to prevent heating of the skin and thus interfering with the measurement, or damaging the tissue. Tens of milliwatts are the desired order of incident light on the skin.
The lamp used here has aluminium, convection cooled, housing and a wall transformer as a power supply. The power input is regulated inside the lamp housing to assure a ~0.4%
stability over the current range. The lamp has a nominal ~. a ...

colour temperature of 2800° K and an average life of 10,000 hours at this colour temperature. The lamp was obtained from Gamma Optronik AB.
It is also possible that a xenon flash lamp could be used instead of the tungsten-halogen lamp. Th.e xenon flash lamp has the ability to be switched on and off at high speed and thus be used as a chopper of the signal. Other advantages are the compact size and the minimal amount of heat generated.
The disadvantage is that the xenon flash lamp output is mainly in the ultraviolet (W) and visible light range.
Chopper To achieve desired wavelengths, two different interference filters were used to select the desired wavelengths from the tungsten halogen lamp.
The wheel (Figure 5.4), which is a metal disc with holes, holds the two interference filters, the filters are selected according to the peak intensity of absorption of water. These band pass filters, which compact and robust, are manufactured by Coherent. One has a centre wavelength at 1460 nm while the other has a centre wavelength at 1940 nm. The third hole on the wheel that has a very small radius allows the light pass through the disc without being filtered. To synchronise the place of the filter with the detected signals, there are holes placed on the periphery of the wheel beside every filter. A
motor whose rate can be controlled drives the wheel: This arrangement acts as a chopper and gives two selective wavelengths with a fixed frequency.
Probe The probe includes an optical fibre and a photo detector.
An optical fibre called Silica - Silica fibre or alternatively known as quartz-quartz fibre, which is the widest used optical p a a.
_ 2~ _ fibre type for low loss light transmission over the widest range of wavelengths, conducts the filtered light from the chopper to the skin. An optical fibre may consist of either one single fibre or many fibres bundled together. Here, a single fibre called HP SIR1000P manufact~r.red by Oxford Electronics, with 1.0 mm diameter core was used. This fibre is suited for NIR region and has operating temperature from -40°C to +85°C.
A uniform and circular patch of light conducted by the optical fibre illuminates the skin. O-H bands in SC absorb one portion of the light and the rest is diffusely scattered by the product into a wide range of angles. The scattered light is detected by the photo detector.
There are two major types of photo detectors to choose, photoconductive and photovoltaic. Speed, sensitivity and spectral response range are the important characteristics for choosing a photo detector. Photoconductive detectors have a high 1/f noise therefore an Indium Gallium Arsenide InGaAs) photovoltaic detector was used here.
The Hamamatsu company provides extended InGaAs photo diodes with an active range between 900 nm and 2100 nm and peak sensitivity at 1950 nm. A small active sensor area was chosen (0.3 mm diameter). The InGaAs photo diode 68422-03 was used. The electronic signals from the detector circuit are of the order of a few tens of microvolts and these signals require amplification. Use of a filter between the detector circuit and the amplifier can eliminate the major problem of the 50 Hz noise interference. The detect~ar circuit converts current from the photo diode to a voltage.
Computer For each wavelength, the voltage pulses - the energy received at the detector - are collected and digitised by a p ~ m plug-in card in a computer. There is a spatial program that can handle these digitised values and give significant information about the measurement. This program calculates average of digitised values under an experiment and gives a value representing the water content in skin. The measurement program is a LabView program, see Figure 9.1. Avoidance of If 50 Hz interference could eliminate the need for a filter before the amplifier.
Measurements In order to reduce noise observed in the 50 Hz area, found in initial trials with a piece of rough aluminum as a reference, a band rejection (notch) (Figure 6.1) filter was used. The filter was applied before the amplifier in the instrument and it was complemented with a low pass filter.
Absorbance measurements using a paper towel (wet and dry) were made. The results are summarized in bar graph of Figure 7.1. The results establish the feasibility of this approach for measuring the moisture content of skin, which can in turn be used in a method for increasing the accuracy of a quantitative in vivo measurement of a mammalian analyte concentration.
The embodiments described herein provide the skilled person with convenient apparatus and method for the determination of water content of skin. The arrangement provides for non-invasive measurement and the NIR light is harmless [Burmeister 1998] and causes no damage to the skin.
The measurement is rapid and its read time could readily be reduced even to one second. For comparison, transepidermal water loss (TEWL) measurements have been :reported to take between 30 and 60 seconds [Nicander 1998]. Further, environmental factors such as temperature and humidity have any effect minimal effect on the measurements.

ay , 1.

The values were found to initially ~recome more positive with time and the drift was about (14-18) mV in 10 minutes.
This may be due to rising temperature. The values shown, however, were stable after about one and a half hours.
Results may be improved through the incorporation of laser diodes which, as described above, would eliminate the need for a mechanical chopper or filters. A laser diode has one wavelength and the light can be chopped on and off rapidly, with concomitant disadvantages of the need for laser diode drivers; and perhaps Pettier cooling.
In an alternative embodiment (Figure 8.1) two photo diodes are used to measure differentially, which requires two sensors per wavelength so four new photo diodes and another lens are included in this embodiment. In the illustrated embodiment, the matched photo diodes can have a max +/- 50 sensitivity (A/W) at the measurement wavelengths (+/- 10 nm) at 25°C. The composition of the extended InGaAs diode increases importance of the position on the wafer. Two matched 68372-01 (D=1 mm, peak sensitivity wavelength at 1950 nm) are used for the measurement at 1940 nm and two matched 68370-O1 (D=1 mm, peak sensitivity at 1550 nm) are used for the 1460 nm measurement.
A particularly useful aspect of the present invention is a method of evaluating water content in skin of a subject animal, typically a mammal, such as a human. The results of the evaluation can be used to calculate a correction factor for use in analyte determination of the subject. A method and apparatus for non-invasive determination of glucose in body fluids is described in international patent application No.
PCT/US 98/02037 published under WO 99/39627 on August 12, 1999 in the name of Dermal Therapy (Barbados) :Inc. The method described includes the use of skin impedance measurements in _u 1Nr making a blood glucose determination and it is for this sort of method that the correction factor can be calculated.
Without being limited by any theory, it is thought that results obtained through such impedance measurements can be improved by subjecting the site of the impedance measurements to prior treatment with an aqueous solution, an inundation step. This inundation step should be conducted for a length of time determined by the pre-existing level of moisture in the skin and can therefore be calculated on the basis of results of the invention disclosed in this specification. By pre-treating the site of measurement for the appropriate length of time, the reproducibility of results obtained by means of the impedance measurements, from measurement to measurement, is enhanced. In other words, it is thought that inconsistencies introduced into blood glucose determinations brought about by variance in skin moisture levels can be compensated for by an inundation step of appropriate length.
One aspect of the present invention is thus a method for increasing the accuracy of a quantitative in vivo measurement of an analyte concentration of a subject, preferably a mammal, the method comprising:
measuring light reflected from a skin surface of the subject, at two wavelengths of said light; and calculating a correction factor based on the measurement.
The correction factor can be a period of time for exposing a skin site of the mammal to an aqueous solution prior to conducting said quantitative in vivo measurement.
In an alternative embodiment, the invention is a method of determining a blood glucose level based on an impedance measurement, wherein the effect of skin hydration on impedance is accounted for through the use of such a correction factor.
Preferably, the skin surface includes said skin site.

Preferably, the mammal is human and said skin si a is located on an arm of t:he human.
Preferably, the quantitative in viva measurement is an impedance measurement.
Preferably, the analyte is glucose of blood within the mammal.
Preferably, the light is infrared light.
Preferably, the infrared light is near-infrared light.
Preferably, a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.
Preferably, the two wavelengths of light differ from each other by at least 300 nm.
Preferably, the first wavelength is at least 1100 nm.
Preferably, the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.
Preferably, the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about.1850 and 2050 nm, more preferably between about y , . >..

1900 and 2050 nm, more preferably between. about 1900 and 200 nm.
In another broad aspect, the invention is a method for calibrating a spectrometer, the method comprising:
measuring a first concentration of an analyte in a sample obtained from the mammal under a first set of conditions;
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light, under said first set of conditions;
measuring a second concentration of the analyte in a sample obtained from the mammal under a second set of conditions;
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light, under said second set of conditions;
calculating a calibration factor based on said first and second concentrations, and said first and second measurements.
In a preferred aspect, one of said first and second set of conditions is preceded by a period of fasting of the mammal.
In another preferred aspect, one of said first and second set of conditions is preceded by ingestion of a predetermined digestible substance.
An apparatus may be provided in which the calculating step is carried out by a pre-programmed computer chip within said spectrometer.
According to another broad aspect, the invention is a method for determining quantitative in vi~ro measurement of a mammalian analyte concentration, the method comprising:
measuring Light reflected from a skin surface of the mammal, at two wavelengths of said light;
calculating a time period based on the measurement;
exposing a skin site of the mammal to an aqueous solution for the calculated period;

a "~ ~ s_~

measuring impedance at said skin site;
determining said analyte concentration based on the measured impedance.
In another broad aspect, the invention is a method of establishing a predetermined skin hydration level at a skin site of a subject, the method comprising:
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;
calculating a time period based on the measurement; and inundating the skin site with an aqueous solution for said time period.
Preferably, the aqueous solution is a saline solution:
Another broad aspect of the invention is a method of establishing a predetermined skin hydration level at a skin site of a subject, the method comprising:
measuring light reflected from a skin surface of the mammal;
at two wavelengths of said light;
calculating a time period based on the measurement; and adjusting the hydration level of the skin at the site based on said calculation.
Preferably, the step of adjusting the hydration level comprises dehydrating the site for said calculated period of.
time.
In another preferred embodiment, the step of adjusting the hydration level comprises exposing the site to an aqueous solution for said calculated period of time.
Another broad aspect of the invention is a method of pre-treating a skin site of a subject prior to measuring impedance at the site, the method comprising:
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;
determining a factor based on the measurement; and > ~ _.

adjusting the hydration level of the skin. according to said factor.
Preferably, determining said factor includes selecting a composition from a predetermined group of compositions for use in said adjusting step.
Yet another broad aspect of the invention is a method of determining a factor for prospective measurement of impedance at a skin site of a subject, the method comprising the steps of measuring light reflected from a skin surface of the subject, at two wavelengths of said light; and calculating a correction factor based on the measurement.
Preferably, the correction factor is a factor for adjusting said impedance measurement.
Another aspect of the invention is thus a method of determining the blood glucose level of a subject that involves measuring impedance of the skin of the subject, wherein the method incorporates thereinto an embodiment of the invention as described above.
In one particular embodiment, therefore, the invention is a method for determining the blood glucose level of a subject, said method comprising':
measuring impedance at a skin site of the subject; and determining the blood glucose level based on the measurement and a correction factor determined according to one of the foregoing methods.
Alternatively, of course, an additional step, e.g., of inundating the skin site according to a method of the invention could be included in determining blood glucose levels.
The invention also provides an apparatus apparatus for determining in vivo moisture content of skin of a mammal. The apparatus includes means for producing light at first and > > ,_t second wavelengths, affixed to a housing so as to direct the light toward a skin site of the mammal when the housing is mounted to the mammal; light detecting means, affixed to the housing so as to detect light produced by the light producing means at the first and second wavelengths and reflected from the skin of the mammal; and means for determining the moisture content based on the light of the first and second wavelengths detected by the light detecting means.
The light producing means is made up of a first laser diode which emits light at said first wavelength and a second laser diode which emits light at said second wavelength in one particular embodiment.
The apparatus can include by a pre-programmed computer chip for determining the moisture content.
The apparatus can include a shroud adapted to substantially preclude extraneous light from reaching the light detecting means when the housing is mounted to the mammal. Preferably, the housing is adapted for mounting to the forearm of a human.

,' ,s . ..

Appendix A - References Anderson R. R., Parrish J. A.: The Optics of Human Skin, The Journal of Investigative Dermatology, 1981:77:1:13-19.
Bokobza L.: Near infrared spectroscopy, J. Near Infrared Spectrosc., 1998:6:3-17.
Burmeister J. J., Arnold M: A.: Spectroscopic Considerations for Noninvasive Blood Glucose Measurements with Near Infrared Spectroscopy, LEOS Newsletter, 1998:12:2:6-9.
Clayton G. B.: Operational amplifiers. (2. Ed.) London:
Butterworth & Co. ISBN 0-408-00370-7. 1979.
Dempsey R. J., Davis D. G., Buice R. G. Jr., and Lodder R. A., Biological and Medical Application of Near-Infrared Spectrometry, Focal Point, 1996:50:2:18A-34A.
Eisenberg D., Kauzmann W.: The structure and properties of water. Oxford: Oxford university press. 1969.
Heise H. M., Marbach R.: Limitation of Infrared Spectroscopy for Non-Invasive Metabolite Monitoring Using the Attenuated Total Reflection Technique, Leos Newsletter. 1999:10:6.
Kumar G., Schmitt J. M.: Optimum wavelengths for measurement of blood hemoglobin content and tissue hydration by NIR
spectrophotometry , Society of photo-optical instrumentation engineers, 1996:2678:442-453.
Martin K., Curtis K.: In Vivo Measurements of Water in Skin by Near-Infrared Reflectance, Applied Spectroscopy, 1998:52:7:1001-1007.
Martin K., Infrared and Raman Studies of Skin and Hair: A
review of cosmetic spectroscopy, The Internet Journal of Vibrational Spectrometry, 1999:3:2.
Nicander I.: Electrical Impedance Related to Experimentally Induced Changes of Human Skin and Oral Mucosa, (thesis), Karolinska Institutet; Stockholm, ISBN 91-628-3097-X. 1998.
Osborne B. G:, Fearn T. and Hindle P. H., Practical NIR
Spectroscopy with Applications in Food and Beverage Analysis. Longman Scientific & Technical, ISBN 0-582-09946-3. 1993.

s e~ . z, Each and every reference cited in this specification is incorporated herein by reference as though reproduced in the specification in its entirety.

ay nv Appendix B -'Word Lisp ATR attenuated total reflecaance biopsy examination of tissue from a living body dermis the deep inner layer of the skin, also called: corium, derma desquamation partial loss of the epidermal basal cells EFL effective focal length feasibility studystudy designed to determine the practicability of a system FP filter Fabry-Perot filter gravimetric analysis of quantities by weight in vitro made to occur outside the body in an artificial environment in vivo occuring or carried out: in the living organism IR infrared keratin fibrous protein that occurs in the outer layer of the skin LASER light amplification by the stimulated emission of radiation LED light-emitting diode Melanin a group of black or dark brown pigments present in skin nano- prefix denoting 10-9 NIR near infrared NIRS near infrared spectroscopy notch filter band reject filter PbS lead sulphide PbSe lead selenide pico- prefix denoting 10-la PS power supply QH quartz halogen SC stratum corneum, the outermost layer of skin SNR signal to noise ratio TEWL transepidermal water loss W ultra violet

Claims

1. A method for increasing the accuracy of a quantitative in vivo measurement of an analyte concentration of a subject, preferably a mammal, the method comprising:
measuring light reflected from a skin surface of the subject, at two wavelengths of said light; and calculating a correction factor based on the measurement.

2. The method of claim 1, wherein the correction factor is a period of time for exposing a skin site of the mammal to an aqueous solution prior to conducting said quantitative in vivo measurement.

3. The method of claim 2 wherein said skin surface includes said skin site.

4. The method of claim 3 wherein the mammal is human and said skin site is located on an arm of the human.

5. The method of any preceding claim wherein said quantitative in vivo measurement comprises an impedance measurement.

6. The method of claim 5 wherein said analyte is glucose of blood within the mammal.

7. The method of any preceding claim wherein said light is infrared light.

8. The method of claim 7 wherein said infrared light is near-infrared light.

9. The method of any preceding claim wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

10. The method of any preceding claim wherein the two wavelengths of light differ from each other by at least 300 nm.

11. The method of claim 9 or claim 10 wherein the first wavelength is at least 1100 nm.

12. The method of any of claims 9 to 11, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between. about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

13. The method of any of claims 9 to l2, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

14. A method for calibrating a spectrometer, the method comprising:
measuring a first concentration of an analyte in a sample obtained from the mammal under a first set of conditions;
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light, under said first set of conditions;
measuring a second concentration of the analyte in a sample obtained from the mammal under a second set of conditions;
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light, under said second set of conditions;
calculating,a calibration factor based on said first and second concentrations, and said first and second measurements.

15. The method of claim 14, wherein one of said first and second set of conditions is preceded by a period of fasting of the mammal.

16. The method of claim 14, wherein one of said first and second set of conditions is preceded by ingestion of a predetermined digestible substance.

17. The method of any of claims 14 to 16 wherein the calculating step is carried out by a pre-programmed computer chip within said spectrometer.

18. The method of any of claims 14 to l7 wherein said sample is a blood sample.

19. The method of any of claims 14 to 18 wherein said analyte is glucose.

20. The method of any of claims 14 to 19 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

21. The method of any of claims 14 to 19 wherein the two wavelengths of light differ from each other by at least 300 nm.

22. The method of claim 20 or claim 21 wherein the first wavelength is at least 1100 nm.

23. The method of any of claims 20 to 22, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between. about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

24. The method of any of claims 20 to 23, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1500 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

25. A method for determining quantitative in vivo measurement of a mammalian analyte concentration, the method comprising:
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;

calculating a time period based on the measurement;
exposing a skin site of the mammal to an aqueous solution for the calculated period;
measuring impedance at said skin site;
determining said analyte concentration based on the measured impedance.

26. The method of claim 25, wherein said skin site is located on an arm of the human.

27. The method of claim 25 or 26 wherein said analyte is glucose.

28. The method of any of claims 25 to 27 wherein said light is infrared light.

29. The method of claim 28 wherein said :infrared light is near-infrared light.

30. The method of any of claims 25 to 29 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

31. The method of any of claims 25 to 30 wherein the two wavelengths of light differ from each other by at least 300 nm.

32. The method of claim 30 or claim 31 wherein the first wavelength is at least 1100 nm.

33. The method of any of claims 30 to 32, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

34. The method of any of claims 30 to 33, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

35. A method of establishing a predetermined skin hydration level at a skin site of a subject, the method comprising:

measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;
calculating a time period based on the measurement; and inundating the skin site with an aqueous solution for said time period.

36. The method of claim 35, wherein said aqueous solution is a saline solution.

37. The method of claim 35 or claim 36, wherein the subject is human, and the skin site is located on an arm thereof.

38. The method of any of claims 35 to 37, wherein said light is infrared light.

39. The method of claim 38, wherein said infrared light is near-infrared light.

40. The method of any of claims 35 to 39 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

41. The method of any of claims 35 to 40 wherein the two wavelengths of light differ from each other by at least 300 nm.

42. The method of claim 40 wherein the first wavelength is at least 1100 nm.

43. The method of claim 42, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

44. The method of claim 42 or 43, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

45. A method of establishing a predetermined skin hydration level at a skin site of a subject, the method comprising:
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;
calculating a time period based on the measurement; and adjusting the hydration level of the skin at the site based on said calculation.

46. The method of claim 45 wherein said step of adjusting the hydration level comprises dehydrating the site for said calculated period of time.

47. The method of claim 45 wherein said step of adjusting the hydration level comprises exposing the site to an aqueous solution for said calculated period of time.

48. The method of claim 47, wherein said aqueous solution is a saline solution.

49. The method of any of claims 45 to 48, wherein the subject is human, and the skin site is located on an arm thereof.

50. The method of any of claims 45 to 49, wherein said light is infrared light.

51. The method of claim 50, wherein said infrared light is near-infrared light.

52. The method of any of claims 50 to 51 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength. of light is in the range of 1500 to 3000 nm.

53. The method of any of claims 50 to 52 wherein the two wavelengths of light differ from each other by at least 300 nm.

54. The method of claim 52 wherein the first wavelength is at least 1100 nm.

55. The method of claim 54, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

56. The method of claim 54 or 55, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm;
more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

57. A method of pre-treating a skin site of a subject prior to measuring impedance at the site, the method comprising:
measuring light reflected from a skin surface of the mammal, at two wavelengths of said light;
determining a factor based on the measurement; and adjusting the hydration level of the skin according to said factor.

58. The method of claim 57, wherein determining said factor includes selecting a composition from a predetermined group of compositions for use in said adjusting step.

59. The method of claim 58, wherein said group of compositions includes a composition capable of hydrating a skin site and a composition capable of dehydrating said skin site.

60. The method of claim 58, wherein said adjusting step includes exposing the skin site to said selected composition.

61. The method of claim 57, wherein said adjusting step includes occluding said site.

62. The method of any of claims 57 to 60, wherein measuring impedance is for determining the blood glucose level of the subject.

63. A method for determining the blood glucose level of a subject, the method comprising a method according to any of 57 to 62, and further comprising the step of measuring impedance at said site.

64. The method of any of claims 57 to 63, wherein said light is infrared light.

65. The method of claim 64, wherein said infrared light is near-infrared light.

66. The method of any of claims 57 to 65 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

67. The method of any of claims 57 to 66 wherein the two wavelengths of light differ from each other by at least 300 nm.

68. The method of claim 66 wherein the first wavelength is at least 1100 nm.

69. The method of claim 68, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

70. The method of claim 68 or 69, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 3.700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

71. A method of determining a factor for prospective measurement of impedance at a skin site of a subject, the method comprising the steps of:
measuring light reflected from a skin surface of the subject, at two wavelengths of said light; and calculating a correction factor based on the measurement.

72. The method of claim 71, wherein said correction factor is a factor for adjusting said impedance measurement.

73. A method for determining the blood glucose level of a subject, said method comprising:
measuring impedance at a skin site of the subject; and determining the blood glucose level based on the measurement and a correction factor determined according claim 300 or 301.

74. The method of any of claims 71 to 73, wherein said light is infrared light.

75. The method of claim 74, wherein said. infrared light is near-infrared light.

76. The method of any of claims 71 to 75 wherein a first said wavelength of light is in the range of 750 to 1500 nm and the second said wavelength of light is in the range of 1500 to 3000 nm.

77. The method of any of claims 71 to 76 wherein the two wavelengths of light differ from each other by at least 300 nm.

78. The method of claim 77 wherein the first wavelength is at least 1100 nm.

79. The method of claim 78, wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between.about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

80. The method of claim 78 or 79, wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm; more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

81. An apparatus for determining in viva moisture content of skin of a mammal, the apparatus comprising:
means for producing light at first and second wavelengths, affixed to a housing so as to direct the light toward a shin site of the mammal when the housing is mounted to the mammal;
light detecting means, affixed to the housing so as to detect light produced by the light producing means at the first and second wavelengths and reflected from the skin of the mammal;
means for determining the moisture content based on the light of the first and second wavelengths detected by the light detecting means.

82. The apparatus of claim 81 wherein said light producing means comprises a first laser diode which emits light at said first wavelength and a second laser diode which emits light at said second wavelength.

83. The apparatus of claim 81 or 82, wherein said first and second wavelengths of light are in the infrared range.

84. The apparatus of claim 83, wherein said first and second wavelengths of light are in the near infrared range.

85. The apparatus of claim 84 wherein a first said wavelength of light is in the range of 750 to 1500 n.m and the second said wavelength of light is in the range of 1500 to 3000 nm.

86. The apparatus of claim 84 or 85 wherein the two wavelengths of light differ from each other by at least 300 nm.

87. The apparatus of any of claims 84 to 86 wherein the first wavelength is at least 1100 nm.

88. The apparatus of any of claims 84 to 87 wherein the first wavelength is between about 1100 and 1600 nm, more preferably between about 1200 and 1600 nm, more preferably between about 1200 and 1550 nm, more preferably between about 1300 and 1550 nm, more preferably between about 1300 and 1500 nm, more preferably between about 1400 and 1500 nm.

89. The apparatus of any of claims 84 to 88 wherein the second wavelength is between about 1600 and 3000 nm, more preferably between about 1600 and 2900 nm, more preferably between about 1600 and 2800 nm, more preferably between about 1600 and 2700 nm, more preferably between about 1650 and 2600 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1650 and 2400 nm, more preferably between about 1650 and 2500 nm, more preferably between about 1700 and 2400 nm, more preferably between about 1700 and 2300 nm, more preferably between about 1750 and 2250 nm, more preferably between about 1800 and 2200 nm, more preferably between about 1800 and 2150, more preferably between about 1850 and 2100 nm, more preferably between about 1850 and 2050 nm, more preferably between about 1900 and 2050 nm, more preferably between about 1900 and 200 nm.

90. The apparatus of any of claims 81 to 89 wherein said means for determining the moisture content is provided by a pre-programmed computer chip.

91. The apparatus of any of claims 81 to 90, further comprising a shroud adapted to substantially preclude extraneous light from reaching said light detecting means when said housing is mounted to the mammal.

92. The apparatus of any of claims 81 to 91, wherein the housing is adapted for mounting to the forearm of a human.