WO1993006774A1

WO1993006774A1 - Method and apparatus for determining hematocrit in blood

Info

Publication number: WO1993006774A1
Application number: PCT/US1992/008358
Authority: WO
Inventors: Eric Fogt; James Kelley
Original assignee: Medtronic, Inc.
Priority date: 1991-10-03
Filing date: 1992-10-01
Publication date: 1993-04-15
Also published as: AU2767692A

Abstract

A method and apparatus for determining the level of hematocrit in blood passing through a blood circuit is disclosed. The method involves repeatedly providing infrared light to the blood in cuvettes (7a, b), measuring the amount reflected, and using the amount measured to calculate the hematocrit level of the blood according to one of two formulas. In one embodiment, the light is at the isobestic wavelength of oxy- and deoxyhemoglobin, and is measured at a single distance from the source. In another embodiment, light at two infrared wavelengths is repeatedly provided and measured, the two wavelengths (usually 810 and 830 mm) being symmetrically disposed about the isobestic wavelength. The two measurements are used to approximate the amount of light reflected at the isobestic point, according to formulas provided, and the approximation is used in the calculation of hematocrit. Apparatus is also disclosed which utilizes LEDs (20, 22, 24) as the light sources, a photodiode (258) for detecting the amount of light, and fiber-optic channels (12, 14, 16, 18) in cables (5) for transmitting it to the blood and from the blood to the photodiode (258). Software is utilized for the approximations and calculations.

Description

METHOD AND APPARATUS FOR DETERMINING HEMATOCRIT IN BLOOD

BACKGROUND OF THE INVENTION Field of the Invention This invention relates to fiber-optic sensors for determining oxygen saturation and hematocrit of the blood as it flows through an extracorporeal blood circuit in which it is oxygenated during bypass surgery and the like.

Description of the Prior Art

Fiber-optic sensors for the determination of oxygen saturation (amount of oxygenated hemoglobin), and/or the determination of hematocrit (amount of red blood cells), in the blood in an extracorporeal circuit or otherwise are disclosed in U.S. Patents 4,444,498, 4,447,150, 4,651,741, 4,745,279, 4,776,340 and others. Many such sensors function by providing light to the blood sample via a fiber-optic cable, and measuring the amount of reflected or transmitted light. Light-emittin diodes (LEDs) have been used in these devices to provide radiation at given wavelengths to the sample via the fiber-optic input cables; other fiber-optic cables have transmitted the light reflected by or transmitted through the sample at those wavelengths to photodiodes to produce a current proportionate to the reflected or transmitted light. In general, the LEDs alternately, or in series, emit light of one wavelength and then another. See Heinemann, U.S. Patent 4,444,498, Lavalee, U.S. Patent 3,647,299, Goldring, U.S.

Patent 4,684,245. The calculation of both hematocrit and oxygen saturation allow calculation of approximately the total oxygen content of the blood.

In sensors for oxygen saturation, often called oximeters, oxygen saturation has been determined using a number of different equations based on input of light at about 665 nm and about 820 nm, such as: sO₂ = A+B(Ratio₀) and sO₂ = A-B (Ratio₀) where Ratio₀ = R₈₂₀/R₆6₅ and A and B are constants. These equations may depend on the configuration of the particular device. 665 nm is a point of large difference between absorption of deoxyhemoglobin and oxyhemoglobin on the reflection curves; 820 nm or thereabouts is an isobestic point at which the absorption is identical for both compounds. Hematocrit has been calculated in such devices from a ratio of the amount of light reflected from the blood at two different distances from the light source; the isobestic point on the oxy and deoxyhemoglobin reflection curves (about 820 nm) is the wavelength suggested because it is not influenced by the oxygen saturation level. Two detectors are used to receive the reflected light at two different distances from the source, all as disclosed in Moran, U.S. Patent 4,776,340, and Schmitt, et al, An Integrated Circuit-Based Optical Sensor for In Vivo Measurement of Blood Oxygenation, Vol . BME-33, No . 2 IEEE: Transactions on Biomedical Engineering, February 1986.

In Steinke, et al., Reflectance Measurements of Hematocrit and Oxyhemoglobin Saturation, American Journal Physiology 1987, pp H147 to H153 and Schmitt, et al., New Methods for Whole Blood Oximetry 1986, Pergamom Press Ltd., hematocrit is similarly determined, but with two sources at different distances from a single detector channel. These systems always, however, require three channels to calculate hematocrit, either two source channels and one detector or vice versa. In the Schmitt article, it is further required that the "near" measurement be taken at a higher wavelength than the "far" one. In addition, Moran teaches that it is not desirable to try to obtain a measurement of hematocrit at the isobestic wavelength, due to the difficulty of obtaining LEDs at the correct wavelengths. In Karkar, U.S. Patent 4,745,279, hematocrit in a blood circuit is determined by measuring light diffused by the blood, the light being provided by an LED source adjacent the blood at the isobestic wavelength. A correction is then made for oxygen saturation, and a look-up table used to determine true hematocrit.

The latter of the two Schmitt articles, New Methods for Whol e Blood Oximetry, 1986, Pergamom Press Ltd. discloses an oxygen saturation measurement made from a device having two identical infrared sources at different distances from the detector, as well as a red light source. Hematocrit (and oxygen saturation) is measured based on reflected light at these wavelengths using a look-up table. However, the Schmitt device would require the separate development and input of a particular look-up table not only for each instrument, but for each blood-contacting unit used, a lengthy and time-consuming process for any device which is manufactured on other than a custom basis, or which uses disposable blood-contacting units.

The above description of art is not intended to constitute an admission that any patent, publication or other information referred to is "prior art" or is enabling with respect to this invention, unless specifically designated as such. In addition, this section should not be construed to mean that a search has been made or that no other pertinent information as defined in 37 C.F.R. § 1.56(a) exists. There are a number of problems, however, associated with the above methods and devices for determining hematocrit and oxygen saturation. In many, hematocrit measurements are simply not as accurate as desired; in others, at least three fiber-optic channels are always required; and in others, there is no method for compensating for LEDs that do not emit at the actual isobestic wavelength. In some, the look-up tables used to calculate hematocrit must be redetermined and reprogrammed into the instrument for each use. Finally, some of the devices are not conveniently set up so that the instrument and fiber-optics can be reused; instead, expensive fiber-optics may be accommodated in disposable units.

Overall, it would be desirable to provide an easily- manufacturable device and a method which provides an easy to accomplish hematocrit measurement at desired accuracy levels.

SUMMARY OF THE INVENTION

The present invention is a method and apparatus for determining hematocrit in blood passing through a blood circuit which avoids many of the problems associated with prior methods and devices.

In a first aspect, the invention is a method for determining hematocrit in blood in a blood circuit. It includes the following steps:

repeatedly providing light at a first infrared wavelength to the blood,

repeatedly measuring the amount of light at the infrared wavelength reflected from the blood at a single distance from the provided light,

repeatedly using the above measurement to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin, and repeatedly calculating hematocrit based on the above approximation of reflected light at the isobestic wavelength. In this aspect of the invention, measurement of reflected light is made at a single distance from the light source, rather than making separate measurements at a near and a far distance.

In s aspect, the first wavelength can be at about the isobestic wavelength of oxy- and deoxyhemoglobin, so that the measurement at the isobestic wavelength can be used directly as the approximation of the reflected light at the isobestic wavelength. Where R_BQ_BESTIC i^s the approximation of reflected light, percent hematocrit is usually calculated according to one of the two following formulas where A, B, C, D, L, and M are constants

and

Hct = Exp (L ·R_ISQBESTIC + M/Ln (R_ISQBESTIC ²) ) In this aspect, in one embodiment, the method also includes the following steps:

repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit; and

repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength at the same distance from the provided light, where the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin. The measurements of infrared light are then used to calculate the approximation of the amount of reflected light at the isobestic wavelength. Preferably the approximation of the reflected light is the average, or the product, of the infrared measurements. In the preferred embodiment, the two wavelengths are 810 and 830 nm.

In another aspect, the invention is another method of determining hematocrit in blood in a blood circuit. It includes the following steps:

repeatedly providing light at a first infrared wavelength to the blood,

repeatedly measuring the amount of light at the first infrared wavelength reflected from the blood,

repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit, and

repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength where the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin,

repeatedly using the measurements of infrared light to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin, and

repeatedly calculating hematocrit based on the approximation of the amount of reflected light at the isobestic wavelength. In this method, a representation of the amount of reflected light at the isobestic wavelength can be made without actually measuring at the specific isobestic wavelength.

In this aspect, the approximation is preferably made according to one of the two following formulas, where R represents the reflected light at the given wavelength:

and

R_ISQBESTIC = R_{FIRST INFRARED} X R_{SECOND INFRARED}.

In this aspect, it is then preferable to calculate hematocrit using one of the two formulas for hematocrit mentioned above.

In another aspect, the invention is an apparatus for determining hematocrit in blood in a blood circuit. The apparatus includes: means for repeatedly providing light at a first infrared wavelength to the blood,

means for repeatedly measuring the amount of light at the infrared wavelength reflected from the blood at a single distance from the provided light.

means for repeatedly using the measurement of infrared light to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin, and means for repeatedly calculating hematocrit based on the approximation of the amount of reflected light at the isobestic wavelength. This apparatus measures the reflected light at a

single distance from the input light, thus avoiding the necessity of measurements at two different distances.

In this aspect, preferably the means for providing light is a light-emitting diode, the means for measuring the amount of light is a photodiode and the means for calculating is comprised of software. Also, the blood is usually passed through a cuvette, the light-emitting diode provides light to the blood through fiber-optic cables, and the photodiode receives light through a fiber-optic cable.

Preferably the first wavelength is at or about the isobestic wavelength of oxy- and deoxyhemoglobin, so that the means for approximating is software which directly uses the measurement of light for the approximation. The software for calculating percent hematocrit usually uses one of the two formulas set forth above. In one embodiment, the apparatus also includes:

means for repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit, and

means for repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength at the same distance from the provided light, wherein the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin. The measurements of infrared light are used for the approximation of the reflected light at the isobestic wavelength.

The means for approximating usually is software which uses the average, or the product, of the measurements of reflected infrared light for the approximation. Most preferably, wavelengths of about 830 nm and about 810 nm are used.

In another aspect, the invention is an apparatus for determining hematocrit in blood in a blood circuit. It includes:

means for repeatedly providing light at a first infrared wavelength to the blood,

means for repeatedly measuring the amount of light at the first infrared wavelength reflected from the blood,

means for repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength where the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin, means for repeatedly using the measurements of infrared light to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin, and

means for repeatedly calculating hematocrit based on the approximation of the amount of reflected light.

In this aspect, preferably the means for providing light are light-emitting diodes, the means for measuring the amount of light is a photodiode and the means for calculating is comprised of software. The blood is preferably passed through a cuvette, the light-emitting diodes provide light to the blood through fiber-optic cables, and the photodiode receives light through a fiber-optic cable.

Usually, the means for approximating is software which calculates the approximation, i.e., R_ISOBESTIC, according to one of the two equations set forth above. The software usually calculates hematocrit according to one of the two formulas also set forth above.

In the above aspects, as well as in other aspects, the present invention provides an advantageous method and apparatus for determining hematocrit which avoids problems associated with various prior methods and devices. BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is an overall illustration of the entire apparatus.

Fig. 2A is a plan view of the end of the fiber-optic cable where it attaches to the cuvette. Fig. 2B is a longitudinal cross-section of the cuvette attached to the cable.

Fig. 3A is a cross section of the bulkhead connector between the cable and the instrument. Fig. 3B is an enlargement of the area where the fiber-optic channels connect and Figs. 3C and 3D are isometric end views of the two parts of the bulkhead connector.

Fig. 4A is a longitudinal cross-section of the cuvette. Fig. 4B is a cross-section of the cuvette taken at lines 4B-4B of Fig. 4A. Fig 4C is a top view of the cuvette and window. Fig. 4D is a cross-section of the calibration cuvette. Fig. 5A is a cross-section of the lock or latch mechanism attaching the cuvette to the main cable; Figs. 5B and 5C are enlargements. Fig. 5D is a side elevation of the latch mechanism.

Figs. 6A and 6B form a system block diagram for the apparatus. Figs. 7A and 7B are flow charts, respectively, of the software for the "Calibrate" mode and the "Run" mode.

Figs. 8A and 8B are tables of partitioned equations for the determination of oxygen saturation. Fig. 8C is a flow chart of software performing the oxygen determinations using the partitioned equations of Fig. 8A, and Fig. 8D is a flow chart of software using the partitioned equations of Fig. 8B.

Fig. 9 is a graph of the reflection curves of deoxygenated and oxygenated hemoglobin showing the isobestic point. DETAILED DESCRIPTION OF THE SPECIFIC EMBODIMENTS

The present method of determining oxygen saturation is accomplished using the oximetric device described here.

I. Overall Fiber-optic Sensor System

The preferred embodiment of the system 1 shown in Fig. 1 includes an instrument or electronic module 3 housing LEDs, photodetectors, and hardware and software. Two non-disposable fiber-optic cables 5 and two cuvettes 7a and 7b, which are placed in the blood stream path, are included so that oxygen saturation and hematocrit in blood passing through the cuvettes in the extracorporeal blood circuit can be measured. A printer (not shown) can be attached to the module or housing 3. One cuvette 7a is located in the venous flow path and one, 7b, is located in the arterial flow path, so oxygen saturation and hematocrit of both can be monitored and displayed. The device is physically and operationally identical for both the arterial and venous paths, except where indicated below; therefore only one of the two paths is described below unless specific description of the other path is necessary.

II. Fiber-optic Cable

Referring to Figs. 2A and 2B, the preferred cable 5 includes four fiber-optic channels. The fiber-optic channels are formed of polymethyl methacrylate with fluorinated polymer cladding obtained from PolyOptical Products, Inc., Santa Ana, California. One is an emitter channel 10 for transmitting 660 nm radiation from a 660 nm LED (20 in Fig. 6A) to the blood sample; a second is an emitter channel 12 transmitting 810 nm infrared radiation from an 810 nm LED (22 in Fig. 6A) in module 3, to the blood sample; a third is an emitter channel 14 transmitting 830 nm IR radiation from an 830 nm LED (24 in Fig. 6A) in module 3 to the, blood sample. A single LED at the true isobestic wavelength (about 820 nm) and a single corresponding channel can be chosen, if available, to avoid the need for both the 810 and 830 nm channels. These emitter channels transmit the radiation through the window of the cuvette 7, discussed below, to the blood sample, illuminating the blood. The fiber-optic cable 5 also includes a fourth channel 16 for receiving reflected radiation at the input wavelengths, and transmitting it to a signal photodetector 258 (shown in Fig. 6A) in the electronic module. This preferred embodiment measures reflected light, as indicated, but it is within the scope of the invention to measure transmitted light instead, in the case of an appropriate device.

At the distal end of the cable, the fiber-optic channels are spaced with the emitter channels equidistant from detector channel 16, as shown in Fig. 2A. The preferred fiber-optic channel size for all the emitter channels is .030 inches in diameter, the detector channel is preferably .040 inches in diameter and the emitter channels are preferably located about .047 inches from detector channel 16. For a particular device, the distance of the emitter channels from the detector channel at the blood interface can be optimized using methods known in the art.

III. Bulkhead Connector

The proximal end of the cable is attached to instrument 3 via a bulkhead attachment 30 shown in cross-section in Figs. 1 and 3A, and in 3B and 3C. A round female unit 32 having interior threads 34 and a key 36 is rotatable with respect to the cable. A male unit 38 with threads 31, mounting unit 7 for mounting in housing 3, and key slot 35 is disposed in module 3. To connect the cable to the module, the male unit is placed within the female unit, key slot to key, and the female unit 32 is rotated to attach the cables to the bulkhead, companion fiber-optic channels not more than .010 inches apart from each other. An enlargement of the connection is shown in Fig. 3B, with a small space 33 between the companion channels.

IV. Cuvette

A cuvette 7a or 7b shown in Figs. 2B, 4A, 4B, and 4C is placed within the extracorporeal blood path for blood flow therethrough during bypass surgery. The blood flow longitudinally through the cuvette through axial cavity 50. The cuvette is molded of polycarbonate for convenient light transmission and is of generally cylindrical shape. Barbed fittings 51 at each end are designed to fit within the tubing of the extracorporeal circuit with a leak-free seal.

V. Latch Connector

When in use, the cuvette is attached to the cable via lock or latch 70 shown in Figs. 2A-B. On cable 5, the lock includes knob 72 rotatable with respect to foot 73 containing the fiber-optic channels. Foot 73 has stops 74 and 76 which limit rotation of knob 72 when they contact extensions 78 and 80. Springs 81 bias foot 73 away from cable 5. Stops 74 and 76 contain key slots or channels 82 and 84, preferably of different sizes. Extensions 78 and 80 together with stops 74 and 76 define an area of rotation of knob 72; the knob can be rotated so that the key slots are located at a position A or a position B as shown. Adjacent the key slots when they are located in position B are lips 86 and 88, each with its own stop 90 and raised lock 92. Cuvette 7 contains keys 94 and 96, each designed to be received in one of slots 82 or 84; space 95 between them is designed to receive foot 72. Each key defines an aperture 98 and flange 97 which can receive a lip 86 or 88.

The cuvette is attached to the cable by sliding the keys into the appropriate key slots when knob 72 is in position A.

Cuvette 7 and foot 73 are then urged toward the cable and the knob rotated so that cuvette flanges 97 pass over locks 92 and reach stops 90 with the knob in position B. Pressure on the cuvette and foot is then released and spring 81 in cooperation with locks 92 firmly hold the channel and cuvette in place with channels 10, 12, 14, 16 abutting window 52 of the cuvette.

VI. System Block Diagram

A system block diagram for the entire device is shown in Figs. 6A and 6B. Bulkhead connector 30 contains three optical channels 250, 252, and 254 to conduct LED illumination to the bulkhead. It also contains one optical channel 256 to conduct reflected signal energy to the signal detector 258 and one optical channel 260 to conduct disconnect signal energy to the disconnect detector 262.

Signal detector 258 is connected to the signal detector amplifier 264 that amplifies the signal in a range usable for the analog-to-digital converter 266. The analog-to-digital converter data is fed to the CPU 270.

Optical channels 250, 252, and 254 are coupled to individual LEDs 20, 22 and 24 of the three wavelengths 660 nm, 810 nm, and 830 nm, respectively. Each LED is energized by its LED driver, shown as one of 274a, b, or c on Fig. 6A. Each LED driver is individually turned on and off by the CPU.

The intensity of each LED is controlled by an individual compensation circuit. Each compensation circuit consists of a compensator detector, one of 276a, b, or c to monitor the LED energy. Each compensator detector is coupled to a compensator detector amplifier, one of 278a, b, or c, which amplifies the signal. Each compensator detector amplifier signal is coupled via a multiplexer, a comparator-integrato and another multiplexer to the appropriate LED driver 274a, b, or c. If the LED energy varies from its set point, this information is fed to the LED driver and the LED driver adjusts the LED energy as necessary to maintain the set point. In operation, individual LED energies are sequentially triggered to illuminate the sample via cable 5. The reflected energy from the sample is transmitted through the detector channel 16 in cable 5 to the signal detector 258. The signal is amplified by the signal detector amplifier 264, converted to digital data via the analog digital converter 266, and read by CPU 270. Preferably, the individual LEDs are triggered so that the detector reads first the "ambient" light received along the channel and then reads the reflected light at the LED wavelength. The "ambient" data is subtracted from the reflected data in the software before correction for calibration.

Referring now to Fig. 6B, the CPU 270 is a standard 8-bit micro-controller with RAM 280, ROM 282, a watch dog timer 284, and an asynchronous serial port 286 for attachment to printer 2 in this case. Additional system components are the power supply 288, an LCD display 290, a test switch 292 and an event switch 294 all connected to the CPU.

VII. Software

A. Calibration

When the device is first used, it is calibrated to compensate for variations in the individual system's optical components; by taking measurements under known conditions, normalization factors can be determined for each LED emitter channel and detector channel combinations. Such factors are d correct the raw readings at each channel for variations inherent in each system.

Calibration is accomplished in the preferred embodiment using a calibration cuvette 500 shown in Fig. 4D. This is a cuvette designed to be attached to the cable unit in the same fashion as a standard cuvette. However, it contains a material designed to provide a known amount of reflected light in the device at the selected wavelengths.

The calibration cuvette does not contain a chamber for blood passage; instead, it contains a well 502 to contain reflective material 504. Preferably, for simplicity of manufacture, it otherwise duplicates the standard cuvette as much as possible.

Also, it terminates in two bases 506 and 508 which are permanently adhesively attached to the instrument module 3 for convenience of use for each calibration. Material 504 is designed to resemble blood in that the amount of light reflected is similar to that of blood. Since the preferred device is designed to operate maximally with blood giving a signal of about 2.0 to 3.0 volts, the preferred calibration standard gives a signal in this range also.

In an attempt to maximize accuracy of the calibration, the device is preferably calibrated using a calibration standard (material 504) providing a voltage just above the maximum reflected voltage level for blood. Using as high a voltage as possible within the desired level minimizes the effect of errors in the calibration reading on the normalization factors which calibration creates. The maximum voltage read from blood in the preferred embodiment at the 660 nm wavelength is about 2.8 volts while the maximum at the 810 and 830 nm wavelengths is about 2.5 volts. In the preferred embodiment,, thus, the calibration standard reflects at a level equivalent to 3.0 volts on the 660 nm channel

and 2.5 volts on the 810 and 830 nm channels. The material used for the calibration includes a light- scattering agent dispersed at desired levels in a support medium. The support matrix is substantially non-light- absorbing at the wavelengths of interest and may be formed of a number of materials such as silicone elastomers (like Dow Coming's Silastic MDX 4-4210), urethanes, epoxy, or other materials. The support matrix is generally chosen for certain characteristics. Specifically, it should cure to a solid fairly quickly without allowing the dispersed materials to settle out. Once cured, its viscosity and other features should be stable over time. The preferred support matrix is epoxy which shows limited degradation over time.

A number of light-scattering materials can be used such as titanium dioxide, Tempera paint pigment, silicon carbide, ferric oxide, etc., which can be evenly dispersed in the matrix. These materials should be generally non-varying over time and, if possible, of a uniform particle size.

In some embodiments, dyes may be used to diminish the reflected light at certain wavelengths. These dyes may be included in the support matrix or may be independently dispersed. Some possible dyes are PSP030, a blue pigment dispersion made by Huls Petrarch, Inc., or Thymol blue made by Fisher Chemical Company. The blue dyes will reduce the amount of reflected red light. The preferred material, specifically, will include about:

59.4% by weight Epoxy resin

35.6% by weight Epoxy hardener

5.0% by weight Titanium dioxide powder

The preferred epoxy is Hexcel Epolite 3310, a 75 Shore D clear epoxy.

The preferred titanium dioxide (or Titanium (IV) Oxide) is a white powder of greater than 99.9% purity in the anatase form.

The mixture is cured in the cavity 502 of calibration cuvette 500 shown in Fig. 4D and is preferably about 1/4 inch thick, or at least thick enough and sufficiently opaque so that ambient light does not enter from the rear. The calibration constants in the preferred device, 3.0 volts for the 660 nm channel, 2.5 volts on the 810 channel, and 2.5 volts on the 830 nm channel are used to calculate normalization factors as follows:

For example, if the voltage read on the 660 channel using the standard is 2.9, the normalization factor for that channel is:

If, for example, on the other two channels, the voltage on the 810 nm channel during calibration is 2.4 and that on the 830 channel is 2.3, the normalization factor for each will be:

and

These factors are stored in RAM as the normalization factors for each channel, and thus readings on each channel are separately calibrated and corrected for variability of the system.

A software flow diagram for the calibration process is shown in Fig. 7A. The CPU first displays "Cal" in the arterial and venous windows 292 to prompt the operator to depress the Test switch 290, and waits for the switch to be held for at least 3 seconds, steps 300, 302 in Fig. 7A. The LED drivers are the directed to energize the LEDs sequentially, step 304.

Data is acquired on the venous and arterial channels at each wavelength as follows. Detector voltages are captured for each wavelength on the venous channel in the hardware capture circuitry described previously. Each captured venous voltage is digitized and stored in RAM in the venous data table, step 304. The same steps are taken for the arterial channel, step 350.

Normalization factors are calculated for each wavelength by the CPU using the acquired data, steps 308, 352. Raw data and the normalization factors are then compared to preset limits for validation, steps 306, 310, 350, 354. If either channel is found to be invalid, the CPU blanks the display. If both channels are invalid, a "Fail" flag is set (and displayed), the "Cal" flag is redisplayed and the procedure must be restarted, steps 314, 358, or problems with the device corrected.

If the venous channel data and normalization factor are found to be valid, an active channel flag is set; the same is done for the arterial channel, step 356. A "Pass" flag is then set and displayed, step 360. If one of the channels is invalid, its display window is blanked.

In use, the CPU uses the normalization factors for each channel to correct raw readings before calculating hematocrit and oxygen saturation via software contained in RAM.

B. Run Mode

After calibration, the device operates in the "Run" mode. All variables, flags and registers are initialized for the run mode and normalization factors are included. The run mode is the normal operational mode for the instrument that is active until the device is powered down. An itemization of one run mode cycle after initial testing (step 400, as described above) is then shown in Fig. 7B.

The detector voltages are captured and processed as described in the calibration section above, steps 402, 404, 450, 452. The output is the venous data table which contains values for each wavelength and the arterial data table, which also contains output for each wavelength.

The venous data is processed as follows. If the venous channel is not active, the procedure is exited, 402, and arterial procedure entered. If active, the venous channel data is compared to acceptable limits, 406. If the venous data is not valid, an error message is displayed in the venous display window, 414, and the procedure is exited. If valid, the data (which has already been corrected for ambient signals) is adjusted for VCO non-linearities.

Finally, the data for each channel is then normalized, using the normalization factors obtained in the calibration mode. This is done by multiplying the received data by the multiplicative scale factor determined during calibration. For example, in the example shown, the received data on the 660 channel will be multiplied by 1.03, on the 810 channel by 1.04, and on the 830 channel by 1.09. Calculation of hematocrit and then oxygen saturation is then completed as described later using the normalized data.

Venous saturation results produced as described above are then displayed, 410, 412. Arterial channel data is similarly processed, steps 462-462, but venous hematocrit is displayed, 416, if the channel is inactive. Otherwise, arteria saturation and hematocrit are displayed, 460. A printer can be used if desired, see 464-468.

VIII. DETERMINATION OF HEMATOCRIT AND OXYGEN SATURATION A. Hematocrit

Where an LED is used which actually produces a wavelength at the isobestic point (about 820 nm) for hemoglobin and deoxyhemoglobin, that measurement is used in the calculation of hematocrit. Specifically, using the normalized data, a reflectance measurement R_ISOBESTIC OR R_I is obtained, as follows for use in the measurement of hematocrit:

R_I ≡ R₈₂₀

Otherwise, LEDs generating wavelengths symmetrically disposed about the isobestic point on the oxy-deoxyhemoglobin curve can be used to generate a representation of the amount of light reflected at the isobestic point. As can be seen on Fig. 9, if an LED generating the exact isobestic wavelength is not available in the device, two LEDs generating wavelengths disposed on either side of the isobestic point can be used¹. The difference between the two curves at the one wavelength should be equal but opposite to the difference between the two curves at the other wavelength.

For example, in the preferred embodiment, LEDs generating wavelengths of 810 nm and 830 nm can be used, although other wavelengths such as 800 and 840 nm can be used. R_ISOBESTIC or R_I can then be represented or approximated as follows:

or

R_I = R_{FIRST INFRARED} X R_{SECOND INFRARED}

lIt should be noted that LEDs used should be screened to determine actual wavelengths emitted, since the actual may not be the same as the labeled wavelength. In the preferred embodiment, then

or

Hematocrit (Hct) can then be calculated by using the R_I value:

This equation, including constants A, B, C and D, was empirically derived by comparing R_I values and their associated known hematocrit values, and is believed to provide very accurate measurements. In the preferred embodiment, the constants have been found to be as follows, using the first equation for R_I provided above. Where the second equation for R, is used, a different set of constants must be generated, as is the case when R_I is measured directly, using a wavelength at about the actual isobestic point on the curve.

A = 19.670148

B = 6.833192

C = 9.833347

D = 0.8262

Under abnormal physiological conditions (such as very unusual flow rates, pH, etc.), it may be desirable to use a modified equation which is believed to be more accurate under such conditions, e.g.:

This equation was developed by including such abnormal test data while generating the hematocrit calculation formula. L and M are constants generated for the device.

B. Oxygen Saturation

Oxygen saturation is determined using a Ratio₀ where:

R₀ = R_INFRARED X R_RED

Specifically, here:

If an actual isobestic wavelength is used instead of the 830 nm LED, then

Oxygen saturation is then calculated as follows, where Hct is the hematocrit value determined above:

E through K are constants determined for the device before manufacture by comparing data from numerous such test devices to actual oxygen saturation figures.

Specifically, in the device shown,

E = 3.462727

F = 36.690131 G = 7.060379

H = 500.693864

I = 0.570215

J = 1425.89362

K = 7708.556902

Different constants will be generated where R_ISQBESTIC is used rather than R₈₃₀.

This approach is believed to provide particularly accurate oxygen saturation data. Not only is the oxygen saturation data corrected for hematocrit variations, but it is also corrected for hematocrit which is measured essentially instantaneously. Furthermore, hematocrit is believed to be calculated very accurately via such actual or approximate reflectance at the isobestic point. Thus, variations in, and errors in the calculation of, the hematocrit level during a bypass procedure does not result in erroneous saturation readings.

Again, to avoid error under abnormal physiological conditions in some embodiments, the following modified calculation of oxygen saturation can be used:

sO₂ = N·Ratio₀ + O/Ratio₀ + P/(Ratio₀ ²) + Q·Ratio₀/Hct + S·Hct + T·Ln(Hct) Determining sO₂ in this fashion is believed to reduce sO₂ error, particularly under abnormal conditions. N through T are constants determined for the individual device.

It was found possible to optimize the calculation of oxygen saturation by partitioning the response curves of the device and including abnormal values. Partitioning was done by both sO₂ level and by hematocrit level to give sixteen formulas, four sO₂ ranges in each of four hematocrit ranges.

Partitioning only by hematocrit level provided four formulas. Fig. 8A contains sixteen particular equations found optimum for the calculation of sO₂ in the preferred embodiment, partitioning by both Hct and sO₂, when used in connection with bovine blood; Fig. 8B lists the four particular equations found optimum when hematocrit levels alone were used. Levels of hematocrit and sO₂ as referred to in the Figures are as follows:

High Hct (level 1) = 33 - 40%

Medium high Hct (level 2) = 26 - 33%

Medium low Hct (level 3) = 21 - 26%

Low Hct (level 4) = 14 - 21%

High sO₂ = 90 - 100%

Medium high sO₂ = 70 - 90%

Medium low sO₂ = 50 - 70%

Low sO₂ = 30 - 50% Software for incorporating the formulas of Figs. 8A and 8B in the device can be prepared by one of ordinary skill in the art. Sample flow charts, however, for such software are provided in Figs. 8C and 8D. It should be noted that in actually utilizing these equations, Ratio₀ is used as an estimate of sO₂ rather than the fully-calculated sO₂ value. Also, in the figures (and the claims) Ratio₀ is sometimes referred to as "Ratio" or "Rat."

This approach is believed to further improve the accuracy of the sO₂ determination, as indicated by a reduction in the standard deviation of the bias with the 4-block partition and a further reduction with the 16-block approach. (Data shows the reduction to be from 2.2 to 1.2 to 1.0.)

IX. Use of the Device

The device is first assembled before use, the second cable unit attached to the instrument using the bulkhead attachment and the cuvettes placed in what will be the extracorporea circuit after it is primed. The cables are attached to the cuvettes using the latches. The device is then calibrated using the calibration cell provided, before priming.

Once the system has been primed and is running, the device can be used to measure hematocrit and oxygen saturation. Using the equations set forth above, the device displays hematocri readings and oxygen saturation readings. A printer can b e activated to print readings if desired. Once the surgery is completed, the cuvette is removed from the circuit and disposed of, and the device readied for another use, with another cuvette, following the procedures described above.

It will be understood that the above description and the illustrations are provided by way of example only, that alternate versions, equivalents, and examples will be apparent to those skilled in the art, and will be within the scope of the invention which is defined by the appended claims. For example, transmitted rather than reflected light can be measured with certain devices.

Claims

WHAT IS CLAIMED IS:

1. A method for determining hematocrit in blood in a blood circuit comprising:

a) repeatedly providing light at a first infrared wavelength to the blood;

b) repeatedly measuring the amount of light at the infrared wavelength reflected from the blood at a single distance from the provided light;

c) repeatedly using the measurement of step b) to approximate the amount of light reflected from the blood a the isobestic wavelength of oxy- and deoxyhemoglobin; and d) repeatedly calculating hematocrit based on the approximation of step c).

2. The method of claim 1 and wherein the first wavelength is at about the isobestic wavelength of oxy- and deoxyhemoglobin, so that the measurement of step b) is used as the approximation of step c).

3. The method of claim 1 and wherein R_ISOBESTIC is the measurement made in step b) and step d) includes calculatin percent hematocrit according to the following formula where A,

B, C, and D are constants:

4. The method of claim 1 and wherein R_ISOBESTIC is the approximation of step c) and step d) includes calculating percent hematocrit according to the following formula, where L and M are constants:

Hct = Exp (L·R_ISOBESTIC + M/Ln (R_ISOBESTIC ²) ) .

5. A method according to claim 1 and further comprising the following steps:

c) repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit; and

f) repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength at the same distance from the provided light, wherein the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin and the measurements of steps b) and f) are used for the approximation of step c).

6. The method of claim 5 and wherein the approximation of step c) is the average of the measurements of steps b) and f).

7. The method of claim 5 wherein the approximation of step c) is the product of the measurements of steps b) and f).

8. The method of claim 5 and wherein the first infrared wavelength is about 830 nm and the second infrared wavelength is about 810 nm.

9. A method for determining hematocrit in blood in a blood circuit comprising:

a) repeatedly providing light to the blood, the light at the isobestic wavelength of oxy- and deoxyhemoglobin;

b) repeatedly measuring the amount of light at the same wavelength reflected from the blood at a single distance from the provided light;

c) repeatedly calculating hematocrit based on the measurement of step b).

10. A method according to claim 9 and wherein step c) includes calculating hematocrit according to the following formula where A, B, C, and D are constants and R_ISOBESTIC is the measurement of step b):

11. A method according to claim 9 and wherein R_ISOBESTIC is the measurement of step b) and step c) includes calculating percent hematocrit according to the following formula, where L and M ar constants:

Hct = Exp (L ·R_ISOBESTIC + M/Ln (R_ISOBESTIC ²) ) .

12. A method for determining hematocrit in blood in a blood circuit comprising:

a) repeatedly providing light at a first infrared wavelength to the blood:

b) repeatedly measuring the amount of light at the first infrared wavelength reflected from the blood;

d) repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength wherein the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin;

e) repeatedly using the measurements of steps b) and d) to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin; and

f) repeatedly calculating hematocrit based on the approximation of step e).

13. A method according to claim 12 and wherein the approximation of step e) is calculated according to a formula selected from the following group, where R represents the reflected light at th given wavelength:

and

R_ISOBESTIC = R_{FIRST INFRARED} X R_{SECOND INFRARED}.

14. A method according to claim 13 and wherein step f) include the step of calculating hematocrit according to a formula selecte from the following group, where A, B, C, D, L, and M are constants:

and

Hct = Exp (L ·R_ISQBESTIC + M/Ln (R_ISOBESTIC ²) ) .

15. Apparatus for determining hematocrit in blood in a blood circuit comprising:

means for repeatedly providing light at a first infrared wavelength to the blood;

means for repeatedly measuring the amount of light at the infrared wavelength reflected from the blood at a single distance from the provided light;

means for repeatedly using the measurement of infrared light to approximate the amount of light reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin; and

means for repeatedly calculating hematocrit based on the approximation of reflection at the isobestic wavelength.

16. Apparatus according to claim 15 and wherein the means for providing light is a light-emitting diode, the means for measuring the amount of light is a photodiode and the means for calculating is comprised of software.

17. Apparatus according to claim 16 and wherein the blood is passed through a cuvette, the light-emitting diode provides light to the blood through fiber-optic cables, and the photodiode receives light through a fiber-optic cable.

18. Apparatus of claim 17 and wherein the first wavelength is at about the isobestic wavelength of oxy- and deoxyhemoglobin, so that the means for approximating uses the measurement of light as the approximation.

19. The apparatus of claim 17 and wherein the software for calculating percent hematocrit comprises the following formula where A, B, C, and D are constants and R_ISOBESTIC is the approximation of the reflection at the isobestic wavelength:

20. Apparatus of claim 17 and wherein the software includes means for calculating percent hematocrit according to the following formula, where L and M are constants and R_ISOBESTIC is the approximation of the reflected light at the isobestic wavelength:

Hct = Exp (L·R_ISOBESTIC + M/Ln(R_ISOBESTIC ²)).

21. Apparatus according to claim 17 and further comprising:

means for repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit; and means for repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength at the same distance from the provided light, wherein the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of . oxy- and deoxyhemoglobin and the measurements thereof are used to determine the approximation of the reflected light at the isobestic wavelength.

22. The apparatus of claim 21 and wherein the means for approximating is software which uses the average of the measurements of the reflected first and second wavelengths of infrared light.

23. The apparatus of claim 21 wherein the means for approximating is software which uses the product of the measurements of the first and second wavelengths of infrared light.

24. The method of claim 21 and wherein the first infrared wavelength is about 830 nm and the second infrared wavelength is about 810 nm.

25. Apparatus for determining hematocrit in blood in a blood circuit comprising:

means for repeatedly providing light to the blood, the light at the isobestic wavelength of oxy- and deoxyhemoglobin;

means for repeatedly measuring the amount of light at the same wavelength reflected from the blood at a single distance from the provided light;

means for repeatedly calculating hematocrit based on the measurement of light.

26. Apparatus according to claim 25 and wherein the means for providing light is a light-emitting diode, the means for measuring the amount of light is a photodiode and the means for calculating is comprised of software.

27. Apparatus according to claim 25 and wherein the blood is passed through a cuvette, the light-emitting diode provides light to the blood through fiber-optic cables, and the photodiode receives light through a fiber-optic cable.

28. Apparatus according to claim 27 and wherein the means for calculating hematocrit uses the following formula where A, B, C, and D are constants and R_ISOBESTIC is the measurement of light:

29. Apparatus according to claim 27 and wherein the software uses the following formula, wherein L and M are constants and R_ISOBESTIC is the measurement of light :

Hct - Exp (L ·R_ISOBESTIC + M/Ln (R_ISOBESTIC ²) ) .

30. Apparatus for determining hematocrit in blood in a blood circuit comprising:

means for repeatedly measuring the amount of light at the first infrared wavelength reflected from the blood;

means for repeatedly providing light at a second infrared wavelength to the blood as it passes through the blood circuit; and means for repeatedly measuring the amount of light reflected from the blood at the second infrared wavelength wherein the first and second wavelengths are substantially symmetrically disposed about the isobestic wavelength of oxy- and deoxyhemoglobin;

means for repeatedly using the measurements of steps b) and d) to approximate the amount of light at the infrared wavelength reflected from the blood at the isobestic wavelength of oxy- and deoxyhemoglobin; and means for repeatedly calculating hematocrit based on the approximation of the amount of reflected light.

31. Apparatus according to claim 30 and wherein the means for providing light are light-emitting diodes, the means for measuring the amount of light is a photodiode and the means for calculating is comprised of software.

32. Apparatus according to claim 30 and wherein the blood is passed through a cuvette, the light-emitting diodes provide light to the blood through fiber-optic cables, and the photodiode receives light through a fiber-optic cable.

33. Apparatus according to claim 32 and wherein the means for approximating is software which uses a formula selected from the following group, where R represents the reflected light at the given wavelength:

and

R_ISOBESTIC = R_{FIRST INFRARED} X R_{SECOND INFRARED}.

34. Apparatus according to claim 33 and wherein the software calculates hematocrit according to a formula selected from the following group, where A, B, C, D, L, and M are constants:

and

Hct = Exp (L·R_ISOBESTIC + M/Ln (R_ISOBESTIC ²))