DEVICE FOR MONITORING ARTERIAL OXYGEN SATURATION

Abstract

The present invention concerns an optical based pulse oximetry device comprising: first, second and third light emitting means, for placement on the skin surface of a body part to inject light in a tissue of said part, the wavelengths of the light emitted by said second and third means being different from each otherlight detecting means located at a relatively short distance from said first light emitting means and at relatively long distance from said second light emitting means and said third light emitting means, for collecting at the skin surface light of said emitting means having travelled through said tissue,first computing means for denoising the output signals of said long distance light detecting means from the output signals of said short distance light detecting means, andsecond computing means for deriving oximetry measurements from the denoised output signals of said long distance light detecting means.

Description

BACKGROUND OF THE INVENTION

1) Field of the Invention

The present invention relates to optical-based pulse oximetry. It concerns, more particularly, a pulse oximetry device for monitoring the oxygen saturation (the so called SpO2) of the haemoglobin in arterial blood.

One very interesting application of the invention is the help of subjects requiring continuous SpO2 monitoring, such as, for example, persons suffering from sleep disturbances, neonates, persons having aerospace and aviation activities, alpinists, high altitude sportsmen.

2) Description of Related Art

Since the early works of T. Aoyagi, the principles of pulse oximetry have been established (J. G. Webster, Design of Pulse Oximeters, Institute of Physics Publishing, 1997).1]. Two contrasting wavelength lights (e.g. λ_r=660 nm and λ_ir=940 nm) are injected in a tissue and a reflected or transmitted part of the photons is further recuperated at the skin surface. The changes in light absorption occurred through the pulsated vascular bed are analysed by means of the Beer-Lambert law. According to this law, the intensity l of light recuperated at the skin surface can be characterized by the expression I=I₀e^αλd, where I₀ denotes the baseline intensity of light and e^αλd models the vascular bed absorption which depends on the absorption index α_λ at the wavelength λ and the vascular bed thickness d.

Due to cardiac activity, the thickness of the vascular bed continuously evolves (d=d₀+Δd(t)) and so does I=I(t). By identifying a characteristic cardiac pattern in both I_r and I_ir, an estimation of the ratio α_r/α_ir can be obtained. Hence, the relative content of oxygenated haemoglobin in the arterial tree is derived by means of an empirically calculated calibration look-up table.

Classical pulse oximeters, one example of which is described in the above-mentioned publication, require the cardiac pattern to be continuously identified and tracked. The apparition of the cardiac activity in the optical intensity is detected by a photo-plethysmograph. The amount of absorbed light correlates with the pulsation of arterial blood, and thus, to the cardiac activity.

In the state-of-the-art, two types of SpO2 probes are currently used, namely reflectance and transmission probes. Both methods are based on the placement of two light sources (LED) and a light detector (photodiode) on the skin surface.

In transmission probes, the optical elements are located on opposite sides of a body part. This configuration assures an easy detection of pulsatile patterns but limits considerably the areas of the body where it can be used: finger-tip, ear-lobes and toe.

In reflectance probes, both optical elements are placed at the same plane of a body surface. The recuperated light is, in this case, backscattered in the body and collected at the skin surface. This configuration virtually allows locating the SpO2 probe at any body placement but creates a severe limitation on its ambulatory use. The probe design must eliminate the possibility of direct light passing from the optical source to the photo-detector (cross-talk or optical shunt). Up-to-date, this limitation has been solved either by glue-fixing the probe to the skin or by means of vacuum techniques. An alternative approach is to further separate the optical components. Hence, the probability of cross-talk is considerably reduced. However, due to the enlarged light-path, a drastic decrease of the received light power is obtained and the detection of pulsatile light becomes troublesome. Some manufacturers have proposed the use of the ECG as an additional recording to overcome such limitations.

The WO 95/12349 publication discloses a pulse oximetry device comprising first, second and third light sources, for placement on the skin surface, light detectors located at a relatively short distance from the first light source and at relatively long distance from the second and third light, and computing means performing a statistical analysis of the noise contributions of the output signals of the long and short distance light detectors for deriving more accurate oximetry measurements.

A disadvantage of this method is that it requires that the light intensities measured at the long and short distances depict enough quality to be used in the computation. Two possibly wrong indications may, therefore, if they are combined, lead to a completely wrong oximetry measurement.

SUMMARY OF THE INVENTION

It is an object of this invention to provide a device for monitoring arterial oxygen saturation that does not suffer from the above mentioned disadvantages.

It is another object of this invention to provide a device for monitoring arterial oxygen saturation that extends the use of reflectance optical-probes to any body location by reducing fixation constrains. Even more, the method overcomes the requirement of an auxiliary ECG recording and restricts the probe to an optical-only-sensor.

These objects are attained according to the invention by providing an optical based pulse oximetry device comprising:

• • first, second and third light emitting means, for placement on the skin surface of a body part of a person to inject light in a tissue of said part, the wavelengths of the light emitted by said second and third means being different from each other,
  • first light detecting means for collecting, at the skin surface, light from said first emitting means having travelled through said tissue,
  • second and third light detecting means for collecting, at the skin surface, respectively light from said second and third emitting means having travelled through said tissue,
  • said first detecting means being located at a shorter distance from said first emitting means than the distance separating said second and third detecting means from said second and third emitting means, and delivering shorter distance output signals representative of the cardiac activity of the person,
  • said second and third detecting means being located at a longer distance from said second and third emitting means than the distance separating said first detecting means from said first emitting means, and delivering longer distance output signals,
  • first computing means for denoising said longer distance output signals by using said shorter distance output signals, and
  • second computing means for deriving oximetry measurements from said denoised longer distance output signals.

In other words, the device of the invention derives an oximetry measurement from only the long distance signals, able to provide a more accurate indication than the short distance signals, which are simply used, as synchronisation (triggering) signals, to denoise the long distance signals. The risk resulting from a possibly double wrong source of information for the final computation is therefore eliminated. This approach is not rendered obvious by the teaching of the already cited WO 95/12349 publication.

According to a first preferred embodiment of the invention, said first computing means is programmed to detect the temporal positions of every maximum of the output signals of said short distance light detecting means, then to perform, from the sequence of the detected maximum positions, a triggered averaging of the output signals of said long distance light detecting means.

According to a second preferred embodiment of the invention, said first computing means is programmed to estimate a representation of the spectral distribution of the output signals of said short distance light detecting means, then to perform, from said estimated representation, the restoring of the output signals of said long distance light detecting means.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features and objects of the present invention will become more apparent by reference to the following description taken in conjunction with the attached drawings, in which:

FIGS. 1 and 4 are block diagrams of two preferred embodiments of an optical based pulse oximetry device according to this invention; and

FIGS. 2 and 3 show two examples of placement of the light emitting means and the light detecting means at the skin surface.

DETAILED DESCRIPTION OF THE INVENTION

When performing reflectance pulse oximetry, two main reasons justify the increase of the physical separation between optical parts (LEDs and photo-diode).

• • 1. In any pulse oximetry probe, the relative pulse amplitude is a good indicator of the quality of the probe placement. This quality factor, usually depicted as Perfusion Index (PI), is also interpreted as a quantification of the width of the vascular bed traversed by a light beam. It was demonstrated (Y. Mendelson, Noninvasive Pulse Oximetry Utilizing Skin Reflectance Photoplethysmography, IEEE Transactions on Biomedical Engineering, Vol 35, No 10, 1988) that, given a reflectance probe, the PI is linearly increasing with the increase of the physical separation between the optical parts.
  • 2. The probability of direct-light being short-cut from the light sources to the light detector by scattering in the outer part of the skin or/and successive reflection in the probe-skin interface is reduced when both optical elements are dispersed. Due to the reduced light short-cut, probe design can be simplified (no glue fixing is required anymore).

However, by increasing the distance between the optical parts, the absolute intensity of received light at the light detector is exponentially decreased and, thus, the quantification of the pulsatile signal becomes problematic, compromising the feasibility of successfully identifying cardiac activity.

In the state-of-the-art reflectance probes, the facts here exposed have imposed a trade-off between:

• • Increasing the physical separation of optical elements, thus reducing cross-talk and increasing the Perfusion Index (PI).
  • Assuring enough light intensity at the photo-detector.

This trade-off has historically forced transmittance probes to include severe fixing mechanism such as glue or vacuum approaches, as described in the already mentioned J. G. Webster publication or by V. Konig (Reflectance Pulse Oximetry—Principles and Obstetric Application in the Zurich System, Journal of Clinical Monitoring and Computing 14:403-412, 1998).

The following table summarizes the advantages and disadvantages of near and far-field photo-plethysmography:

	Received	Perfusion	Cross-
Separation	Light Intensity	Index (PI)	talk	Utility
Near	Excellent	Poor	Likely	Easy detection of
				cardiac activity
Far	Poor	Excellent	Unlikely	Reliable pulse
				oximetry
				estimations

The present invention merges the advantages of both near and far field photo-plethysmography in a single method. As shown in FIG. 1, the invention consists in combining far and near photo-plethysmographs so that:

• • a near-field photo-plethysmograph allows the continuous tracking/detection of cardiac activity;
  • a far-field photo-plethysmograph performs pulse oximetry measurements on the basis of estimated cardiac activity information.

According to the invention, a near-field reflectance photo-plethysmograph and a far-field reflectance photo-plethysmograph are merged in a unique device comprising:

• • for the near-field function, a first light source 10, which can be a LED emitting in the infra-red range at 940 nm, a first light detector 11, such as a photo-diode, located to receive light from the source, and a first analog-to-digital converter (ADC) 12 connected at the output of the light detector;
  • for the far-field function, a second light source 13, such as a LED, emitting in the infra-red range at 940 nm, a second light detector 14, such as a photo-diode, located to receive light from source 13, a second analog-to-digital converter (ADC) 15 connected at the output light detector 14, a third light source 16, such as a LED, emitting in the red range at 660 nm, a third light detector 17, such as a photo-diode, located to receive light from source 16, and a third analog-to-digital converter (ADC) 18 connected at the output of light detector 17;
  • a microprocessor 19 connected at the outputs of analog-to-digital converters 12, 15 and 18; and
  • a display device 20 connected at the output of microprocessor 19.

The above-mentioned wavelength values of 660 and 940 nm are just given as examples. More generally, these wavelengths must be in the visible infra-red region, i.e comprised between 400 and 2000 nm, and be different from each other.

As shown in FIG. 2, light sources 10-13-16 and light detectors 11-14-17 are positioned at the surface of the skin S of a body part. The light detectors are at the same location. Near-field light source 10 is at a shorter distance from the detectors than far-field light sources 13-16, located at the same place.

Typically, the separation between the near-field light source and the light detectors is between 4 and 10 mm, whereas the separation between the far-field light sources and the light detectors is between 10 and 50 mm.

FIG. 2 shows that the light collected by the detectors has travelled in tissue T longer and deeper for the far-field beam F than for the near-field beam N.

The above described structure is a simplified presentation of the device of the invention. Needless to mention that a single light detector and a single analog-to-digital converter can also be used in association with time-sharing control means adapted to apply to microprocessor 19 data corresponding respectively to the three light sources 10, 13 and 16.

According to the present invention, the light sources and the light detectors can be arranged at the skin surface in many different configurations, the only rule to respect being:

• • to collect a light beam having travelled over a short distance in the body, and
  • to collect two light beams of different wavelengths in the visible infra-red region having travelled over a longer distance in the body.

Thus, for example, the three light sources can be located at the same place, with a near-field detector at short distance and far-field detectors at longer distance.

Another example is to have a plurality of light sources distributed around far-field detectors, with a near-field detector located at a shorter distance from one of the sources.

FIG. 3 shows, as a further example (with the same reference letters as FIG. 2), that the device of the invention can be arranged around the finger of a person. In that case, light sources 10-13-16 are located at the same place, near-field detector 11 is near the sources and far-field detectors 14-17 stand opposite to the light sources.

Microprocessor 19 has the following two functions:

Stage 21

• • Due to the increased distance between light sources 13-16 and far light detectors 14-17, the digital signals provided by far-field analog-to-digital converters 15 and 18 are noise polluted and render very difficult a reliable identification of the cardiac activity. But the reduced distance separating light source 10 and near light detector 11 assures enough received light intensity and provides a much better identification of the cardiac activity. In stage 21, the near-field signals are used, therefore, to base the pulse oximetry measurements on an improved far-field information.

Stage 22

• • The signals provided by stage 21 are finally used for conventional pulse oximetry calculations.

As shown in FIG. 1, the digital output of near-field ADC 12 is first applied to a band-pass filter 23, such as a Chebyshev filter Type 1, 3^rd order, having a band-pass of 0.5 to 3.5 Hz. Knowing that the useful portion of the signal corresponds to the normal, around 1 Hz, cardiac frequency of a person, this filter eliminates the portions of the signal which are outside the 0.5-3.5 Hz range.

Similarly, the digital outputs of infra-red far-field ADC 15 and of red far-field ADC 18 are first applied respectively to band-pass filters 24 and 25, identical to band-pass filter 23.

In addition, the digital outputs of infra-red far-field ADC 15 and of red far-field ADC 18 are applied respectively to identical low-pass filters 26 and 27, such as Butterworth filters, 2^nd order, which have the function to eliminate the portion of the received signals above 0.2 Hz. The remaining portion of the signals are taken respectively as the DC-infra-red (DC_ired) and the DC-red (DC_red) components of the far-field signals.

The operation shown in 28 is the detection of the temporal position of every maximum of the signal delivered by band-pass filter 23. The sequence of the maximum position is then used to perform, respectively in 29 and 30, a triggered averaging of the infra-red and red far-field signals produced by band-pass filters 24 and 25. The triggered averaging is performed in a similar way to that described in the already mentioned publication of J. G. Webster. The triggered averaged signals resulting from operations 29 and 30 are taken respectively as the AC-infra-red (AC_ired) and the AC-red (AC_red) components of the far-field signals.

Finally, in stage 22, the DC_ired, DC_red, AC_ired and AC_red signals are used to perform classical pulse oximetry calculations 31, as described by J. G. Webster. The results of the calculations are displayed by device 20 connected at the output of microprocessor 19.

Reference is made, now, to FIG. 4 which presents another method for obtaining pulse oximetry measurements from the signals delivered by band-pass filters 23, 24 and 25 and by low-pass filters 26 and 27. The elements common to the device of FIG. 1 are designated by the same references

As shown in FIG. 4, the near-field signals produced by band-pass filter 23 are used, in 32, to estimate a a-priori representation of the spectral distribution of the cardiac activity, as disclosed, for example, in the publication of D. G. Manolakis, Statistical and Adaptive Signal Processing, McGraw-Hill Higher Education, 2000.

Then, the estimated representation of the spectral distribution of the cardiac activity is used, respectively in 33 and 34, to denoise and/or restore the corrupted infra-red and red far-field signals produced by band-pass filters 24 and 25. The technique used is described, for example, in the already mentioned publication of D. G. Manolakis. The restored signals resulting from operations 33 and 34 are taken respectively as the AC-infra-red (AC_ired) and the AC-red (AC_red) components of the far-field signals. They are finally used to perform the classical pulse oximetry calculations 31.

The present invention can be used in many optical-based pulse oximetry applications. For example, a probe carrying the light sources and the light detectors can be placed:

• • in a head band, the frontal bone acting as reflectance surface;
  • in a mask, the maxillary bone acting as reflectance surface;
  • in a chest-belt, the manubrium acting as reflectance surface;
  • around a finger;
  • around the leg or arm of a neonate;
  • as a ear-phone.

Claims

1. An optical based pulse oximetry device comprising: first, second and third light emitting means, for placement on the skin surface of a body part of a person to inject light in a tissue of said part, the wavelengths of the light emitted by said second and third means being different from each other,first light detecting means for collecting, at the skin surface, light from said first emitting means having travelled through said tissue,second and third light detecting means for collecting, at the skin surface, respectively light from said second and third emitting means having travelled through said tissue,said first detecting means being located at a shorter distance from said first emitting means than the distance separating said second and third detecting means from said second and third emitting means, and delivering shorter distance output signals representative of the cardiac activity of the person,said second and third detecting means being located at a longer distance from said second and third emitting means than the distance separating said first detecting means from said first emitting means, and delivering longer distance output signals,first computing means for denoising said longer distance output signals by using said shorter distance output signals, andsecond computing means for deriving oximetry measurements from said denoised longer distance output signals.

2. The device of claim 1, wherein the wavelength of the light of said second and third light emitting means is in the visible infra-red region.

3. The device of claim 1, wherein said shorter distance is comprised between 4 and 10 mm and said longer distance is comprised between 10 and 50 mm.

4. The device of claim 1, further comprising: bandpass filters connected between said light detecting means and said first computing means, andlowpass fiters connected between said longer distance light detecting means and said second computing means.

5. The device of claim 4, wherein said bandpass filters eliminate the portions of the received signals which are outside the 0.5-3.5 Hz range.

6. The device of claim 4, wherein said lowpass filters eliminate the portions of the received signals which are above 0.2 Hz.

7. The device of claim 1, wherein said first computing means is programmed to detect the temporal positions of every maximum of said shorter distance output signals, then to perform, from the sequence of the detected maximum positions, a triggered averaging of said longer distance output signals.

8. The device of claim 1, wherein said first computing means is programmed to estimate a representation of the spectral distribution of said shorter distance output signals, then to perform, from said estimated representation, the restoring of said longer distance output signals.