l 179777 METHOD AND APPARATUS FOR LINEARIZING RATE ASSAY~
FIELD OF THE INVENTION
Thi6 invention relates to a method and appar-atus for determining a variable in a rate reaction. It is particularly applicable to enzyme reactions, wherein an enzyme transforms one or more substrates to produce a detectable change at a rate the value of which is a measure of the enzyme activity. Thus, the variable to be determined is the rate of the reaction.
BACKGROUND OF THE INVENTION
Rate assays, and in particular, enzyme assays, are determined on the basis of the rate of formation or destruction of an observable signal, such as a detec-table density or fluorescence, over time. The signals form response curves, or profiles~ which can be plotted as a function of time. The greater the amount of enzyme activity that is present, the greater the rate of change in the signal. Thus, the important factor of the reac-tion is the rate, instead of, for example, the end-point.
Non-linear responses in such assays are highly undesirable. Enzyme reaction rates tend to be most meaningful if they are uniquely determinable, for example, if they provide a linear response over time.
If the signals produce a non-linear response curve, any measure of the enzyme reaction rate as the slope of the response curve will produce a non-constant value which depends upon the particular point of the curve that is considered.
Unfortunately, many rate reactions, and espec-ially enzyme rate reactions, produce non-linear response curves. For example, the rate curves shown in Fig. 9 of U.S. Patent No. 4,059,405, issued November 22, 1977, are grossly non-linear. Such non-linearity is not easily eliminated, particularly when measured in a test com-position such as a multi-zoned element contalning reagents. Non-linearity is believed to be introduced ln ~97~
part by a number of non-en7ymatic, and ~hus, extraneous, physical processes. For example, there often occurs an 'induction period" in such element6 in which processes extraneous to ~he enzyme reaction take place. Thi8 period may extend, in certain cases, for the full length of the assay.
Notwithstanding the above-noted non-linearity, if one could be confident that the extraneou6 reactions that occur during the induction period were all com-pleted by a predetermined tlme, then the enzyme activity could always be read with confidence within a given period thereafter as a slope of the curve. However, the induction period i6 not con6tant. It appear~ to vary becau6e of a number of factors.
Furthermore, even if induction periods were constant, deletion of the induction period data could 6everely lim~t the available dynamic range of the a6say, that is, the range of conce~tration or enzyme activity measurable by the assay.
Therefore, non-linear re~ponse curves become difficult to use as a basis for rate assays. The opera-tor never can be certain at what time the extraneous processes are all complete. Becau6e of this uncer-tainty, well-defined continuous calibration models can-not be obtained. Without proper calibration, the observable ~ignal6 re~ulting from the enzyme re~ctions within the test compo~ition cannot be accurately and preci6ely converted to enzyme actlvity.
Recently, it has been found that te6t elements prepared as described in commonly owned Canadian Applica-tion S.~. 420,478, filed January 28, 1983, by Sanford et al, entitled "IncorporAtion of Pyridoxal Phosphate in Dry Analytical Elements for the Determination of Enzymes", provide response curves having improved linearity over a broad dynamic range. However, the curves produced by that invention are not totally lin-ear. Even with that recent invention, there has remain-ed, prior to the instant invention, a need to further improve the linearity of the enzymatic response curves.
For example, the presence of only short l~near portions in the response curves creates short time windows for the assay. The shorter the time window, the more likely it is that instrument noise will affect the reading.
SUMMARY OF THE INVENTION
The invention is ba6ed upon the discovery that, in many cases, the curve indicative of the increasing or decreasing detectable signals produced by a rate reaction in a test composition, for example, can be automatically linearized over at least a portlon of the reaction, even when the "raw" data is initially grossly non-linear. The linearizing is achieved by arith-metically operating on the values of the detected signals produced by the test liquid, using the corresponding values that are de~ermined for a reference liquid having a predetermined analyte activity or concentration. Because the resulting data is more linear, the rate of change is easily ascertained thereafter.
More specifically, in accord with one aspect of the invention, there is provided a method for the deter-mination of an analyte activity or concentration ~n atest liquid by the measurement of a detectable signal that increases or decrea~es over time at a rate that corresponds to the analyte activity or concentration.
The method includes the steps of determining at plural times a value representing the detectable signal at each of the times, and ascertaining the rate at which those values increase or decrease. This method is improved by using, as the value-determining step noted above, the steps of a) determining the value that represents the detectable signal at each of the plural time~ for a reference liquid having a predetermined analyte activity or concentration, b) determining the value repre~enting the detectable signal at the plural times for the noted B ~ ~ ~ 77 test liquid, and c) arithmetically operating on the representative value determined in one of steps a) and b) using the respective value determined in the other of steps a) and b), for each of the plurality of times.
In accord with another aspect of the invention, an analyzer is provided, the analyzer including fir6t means for determining at plural times a value represent-ing the aforementioned detectable signal at the noted times, and second means for computing the rate. The analyzer is improved in that it includes a) means for storing certain of such determined values obtained by the first means for a reference liquid having a prede-termined analyte activity or concentration, b) operating means for arithmetically operating on one of i) the determined values for the test liquid as determined by the first means, and ii) the determined values for the reference liquid stored in the storing means, u6ing the other of i) and ii), and c) means for approximating the rate of increase or decrease of the values computed by the operating means.
Thus, it is an advantage of the present inven-tion that a reliable reaction rate, for example, a rate indicative of enzyme activity, is readily ascertainable, even when the response curves of the "raw" detectable signals produced by the test liquid~ being assayed are grossly non-linear.
It is a related advantage of the invention that such reliable reaction rates are determinable, even us-ing multi-layered test elements in which non-enzymatic physical processes can occur.
A further related advantage of the inveotion i8 that greater linearity provides extended dynamic range.
Yet another advantage of the inventlon is that greater linearity permits the use of lon~er and/or variable time windows for the enzyme assay. This in turn reduces the likelihood of a large protein bias, which occurs usually at the early stages of the reaction reduces the likelihood that instrument nois~ will be a significant ~actor, th~reby imp~oving precision.
Other advantages and festures will become apparent upon reference to the following detailed Description of the Preferred Embodiments, when read in light of the attached drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a schematic illustration of an analy-zer constructed in accordance with the invention;
Fig. 2 i8 a ~chem~tic illustration of com-puting means for the analyzer and its interaction with the photodetector of the analyzer;
Figs. 3A and 3B are logic flow charts for the programming of the analyzer computing meansj Figs. 4 and 5 are graphs of the detectable signals produced by one enzyme in the form of reflection density, versus time, as determined before and after, respectively, subtraction of the reflection density values generated by a reference liquid;
Figs. 6-7 and 8-9 are two sets of graphs simi-lar to those of Figs. 4 and 5, for enzyme assays;
Figs. 10 and 11 also are graphs similar to those of Figs. 4 and 5, except that a reference liquid having a high level of enzyme activity was used for the subtraction step;
Figs. 12 and 13 are a pair of graphs similar to those of Figs. 4 and 5, except that the calibrators were prepared using non-human serum.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The hereinafter-detalled preferred embodiments are described particularly in connection with a te6t composition that is preferably a multi-zoned te~t element. In addition, the invention is useful with other test compositions, and not ~ust multi-zoned ele-ments, particularly if such compositions tend to produce non-linear rate reactions.
~ ~7977~
The preferred embodiments are also described in connection with the determination of enzyme activity levels. In addition, the invention is applicable to the determination of any analyte concentrat~on, particularly those that produce a raw signal that varies non-linearly with respect to time.
Most particularly, the preferred embodiments of the invention are described in connection with the sub-traction of the values obtained for either a reference liquid or the test liquid, from the values obtained for the test liquid or the reference liquid, respectively, at each of the same plural times of the reaction. For example, the values for the reference liquid are sub-tracted from the values for the test liquid. In addi-tion, the invention is applicable to any step of arith-metically operating on the values determined for one of the test liquid and reference liquid, using the respective values determined for the other of the liquids. As used herein, "arithmetically operating on the values determlned for one of the ~est liquid and reference liquid, using the respective values determined for the other liquid", means subtracting from, adding to, dividing into, or multiplying by the values deter-mined for one liquid, the values determined for the other liquid.
A variety of test elements constructed to con-duct A rate assay, such as an enzyme assay, is useful with the invention. Preferred are multi-zoned eiements having a plurality of reagents and/or functions that are divided among the zones. Most preferred are elements ln which the zones are separate but contiguous layers, for example, a multi-layered test element as described in U.S. Patent No. 3,992,158, issued on November 16, 1976, or in U.S. Patent No. 4,258,001, issued on March 24, 1981. The test elements of said patents include an uppermost layer that functions to transport the liquid to be tested to the next ad~acent layer or layers. Such uppermost layer optionally includes a reagent for the 1 ~79~7 ~est, for example, one or more substrates for the enzyme. The next ad~acent layer or layer~ preferably include a matrix or binder and remaining reagents for the assay. These remaining reagents include those necessary to produce a detectable signal in increasing amounts, e.g., a reflection density or fluorescence, in response to the reaction of the enzyme. Alternatively, such rea~ents produce a detectable signal which iB
destroyed or converted to a non-detectable species at a rate proportional to the amount of enzyme present. The detectable Bignal iB incluBive of but not limited to, reflectances and luminescent 6ignals such as a fluorescent signal. ~ther detectable signals are use-ful as well, for example, potentiometric signal~. In the discussion which follows, density-producing or den-sity-destroying enzymatic reactions are emphasized as the most preferred reactions.
This invention is particularly useful in assay-ing a dehydrogenase or a dehydrogenase-coupled enzyme reaction, using t~st elements conta~ning NADH as a co-enzyme. For example, the test elements described in the aforesaid Sanford et al application, e.g., for the assay of aspartate aminotransferase (AST) or alanine amino-tranferase (ALT), are useful.
A highly preferred test element comprises, as the uppermost layer, a non-fibrous, isotropically porous spreading layer such as a layer comprising micro-crystalline cellulose, e.g., that which i6 available from FMC Corporation under the trademark "Avicel", a binder such as polyvinylpyrrolidone, a buffer that main-tains an optimum pH for the enzyme, for example, about 7.2, and a substrate for the enzyme being assayed. The ad~acent layer in such an element preferably comprises a hardened gelatin binder, a surfactant, a buffer that maintains an optimum pH and a color-producing or ~ '7 colored species such as NADH. Such layer is dispo~ed on a trsnspsrent support, such as a plastic or glass.
Using test elements constructed aR described above, this invention i~ particularly useful in assaying for lactate dehydrogenase (LDH). In that u6age, the substrate that is incorporated into the uppermost layer is sodlum pyruvate. When LDH encounters this substrate, the enzyme catalyzes the following reaction:
sodium pyruvate + NADH + ~ LDH
Alternatively, the above-noted reagents sodlum pyruvate and NADH are useful if disposed all in one layer--for example, in the uppermost layer or in a single-zone te~t element. Preferably, in such an element, the matr~x i6 the same a8 iB described for the uppermost layer in the previou6 paragraph. For best results in assaying LDH, such an element is used fresh, rather than after extensive storage.
The invention 18 useful in ~ny analyzer con-structed to perform rate a6says, that is, to repeti-tively examine each test compo~ition rapidly 80 that the rate of increase or decrease of the detectable signal of each test composition can be determined.
IJseful analyzers include those described in U.S. Patent No. 4,224,032, issued on September 23, 1980.
Such an analyzer lQ comprises, Fig. l, a test llquid supply means 20, a test element supply means 30, metering device 40, incubator 50, transfer means 45 for transferring a test element from supply means 30 to me~ering device 40 ~nd thence to incubator 50, A
photometer 70, and a computing means lO0. More specifi-cally, the test liquid supply mean6 i8 a rotatable tr~y 22 containing a plurality of dispos~ble containers 24 and means for rotating the tray, not shown. Test ele-ment supply means 30 comprises a rotAtable t~ble 32, ~ 179777 which mounts a plurality of stacked test elements A, B,C and D, and drive means, not shown, for rotating the table so that one or another of the stacks of test ele-ments is presented to an ejector mechanism 34 which feeds them to transfer device 45. In one form of the analyzer, metering device 40 comprises a station con-structed to lift one of said containers 24 to a pres-surizing means, said containers 24 having a dispensing aperture, not shown, which permits the formation of a pendant drop of liquid in response to the pressure gen-erated by the pressurizing means. Transfer means 45 positions a test element under such pressurized contain-er at metering device 40, to contact the drop to the test element.
The transfer means 45 of the afore6aid U.S.
Patent No. 4,224,032 comprises a rotatable arm 90. Arm 90 is provided, not shown, with test element receiving rails, an ejector blade, means for rotating the entire arm, means for reciprocating the blade, and means for raising and lowering arm 90.
Incubator 50 comprises, Fig. 1, a rotor 52 pro-vided with multiple stations each adapted to receive a test element, a drive shaft 54, and rotor drive means 56. The stations are disposed around the circumference of the rotor to mount the elements in planes that are either parallel to drive shaft 54, as shown, or perpen-dicular thereto. Optionally, means are included, not shown, to load and unload rotor 52 "on the fly", that is, without stopping its rotation. Drive means 56 is adapted to rotate rotor 52 at a fixed rate, for example, 12 RPM.
Photometer 70 comprises a light source 72, optics for directing light as a beam 76 into the inter-ior of rotor 52 while it rotates, and a photodetector 78. As is conventional, beam 76 is inclined at 45 to the normal of the test elements, while radiation is ~ ~97~'7 detected therefrom as diffuse radiation reflected along the normal, arrows 80. The to-be-detected radiation i~
reflected to photodetector 78 by, for example, mirror 82 disposed lnside rotor 52. Filters, not shown, are placed, lf desired, either between light source 72 and the test element in rotor 52, and/or between such test element and photodetector 78.
The 6ignal generated by photodetector 78 i6 directed to computing means 100, which function6 with control center 102 to control the operations and cal-culationæ of the analyzer. As i6 conventional, such computing mean6 further include input/output devices, such as keyboard 104 and display 106. Control center 102 features, for example, driver interface boards to convert computer æignals to æignals that control the motoræ of the various moving componentæ of the analyzer.
Still other analyzeræ are capable of cRrrying out the instant invention. For example, metering device 40 and containers 24 can be altered to be thoæe described in U.S. Patent No. 4,287,155, iæ6ued on September 1, 1981, wherein liquid is aspirated from con-tainers 24 into a pipette, and then dispen6ed onto the ~est element. In yet another, highly preferred embodiment, drive means 56 of incubator 50 is modified so as to stop the rotor every time a test element is to be loaded or unloaded, rotor 52 being controlled to stop so that beam 76 does not illuminate a test element during such stop-pa~e (to prevent pllotolysis) Other photometer constructions sre al80 useful in the analyzer. For example, uæeful photometers include tho6e constructed BO a6 to rotate independently of the rotation of rotor 52, for example, as described in U.S. Patent No. 4,119,381, is6ued on October 10, 1978.
l 179777 If fluorimetric signals are to be detected, the photometer is replaced with a conventional fluorimeter and fluorimetric signals are detected rather than reflection densities.
In accord with one aspect of the invention, computing means 100 is programmed to collect, via an amplifier 90 and analog-to-digital converter 92, Fig. 2, the values representing, for the preferred embod~ment, the reflection densities of the test elements containing a test liquid, at the times the elements are scanned by the photometer, and to subtract therefrom the values obtained at the same times, representing the reflection densities of a test element containing a reference liquid having a predetermined enzyme activity. As used herein, the term "reference liquid" means a liquid having any predetermined enzyme activity level including zero activity, because even high activity levels are useful in the reference liquid. As a result, computing means 100 is adapted to approximate from the difference values obtained by subtraction, the rate of change of the reflection densities produced by the test liquids, which rate represents the enzyme activity under investigation.
To this end, any suitably programmed digital computer is useful in the invention. Preferred are analyzers containing a programmable microprocessor 100', Fig. 2, which comprises a central processing unit 200, for example, an Intel 8086 chip, and memory units 202 comprising one or more RAM's 204 and optionally one or more E PROMI 6 208.
In accordance with the preferred process of the invention, the values determined as representing the detectable signals generated by the test liquids in the test elements at the plural times they are scanned by the photometer, are processed to subtract therefrom the values representing the detectable signals, at the same ~7~7~7 corresponding times, of an identical test element con-taining a reference liquid. It is the dlfference value obtained from this subtraction, that is used to establish the rate of change of the signal. The deter-mined values are preferably reflection densities orreflectances, where density ~R = -log10 (reflec-tance). In carrying out this process, the "raw" density calculated from the reflectance actually detected at each scanning time is used, or, a plurality of readings, closely-spaced in time, is taken during each scanning time, from which a mean value and the corresponding density are calculated. Such a technique i8 deBCribed, for example, in U.S. Patent No. 4,055,752. To u6e such latter technique, photometer 70 provides multiple readings each time a test element is scanned. Esch of these readings represents the density of a fraction of the total density-producing area of the element exposed to beam 76. More specifically, the procedure is to ~can at time tl the density-producing portion of the test element, and integrate the light detected for that area. This integrated signal is amplified at 90 and sent to the A/D converter 92, Fig. 2, where a plurality of digital values are obtained from the integrated signal. These digital values are averaged to obtain the mean density value for time tl. At time t2, which is, for example, 5 seconds later, the process is repeated. The programming of computer means 100 or 100' to determine such values, is conventional.
If mean values are used to obtain reflectances or reflection density values generated by the test liquid, preferably mean values are similarly ascertained in order to determine reflectances or reflectlon density values generated by the reference llquid.
To convert reflectance signals to reflection densities (abbreviated, "DR"), the following equation is used by the computing means, disregarding a correction factor which can be ignored:
~9~ 7 l)DR = -logl0 [(SAlD-~A/D)/(RAlD ~A/D)]
where SA/D is the A to D conversion value taken for the light intensity of either the reference liquid or the test liquid, RA/D is such A to D conversion value for a white reference that is part of the analyzer, and ~A/D is the A to D conversion value for the light intensity produced by a dark or black reference that is psrt of the snalyzer.
In some enzyme chemistries, for example, y-glutamyltransferase and alkaline phosphatase, trans-mission densities (DT) are preferred. In that case, a separate branching routine iB followed by the computing means to convert DR to DT. Such conversion is described in 24 Clin. Chem., pp. 1335-1342, and espec-15 ially p. 1340, (1978), the details of which are expressly incorporated herein by reference, except that the conversion equation set forth in the article should be
2) DT = -0.194 + 0.469 DR + 0.422/(1~1.179e3-379~).
Optionally, and particularly if the detectable signals determined for the reference liquid are greater than those of the test liquid, a constant value is added back to each of the difference values to convert all readings into positive values.
I have discovered that, once the subtraction of the reference liquid DR (or DT) values has occurred, for many chemistries the resulting values are linear over extended periods of time, so that the constant rate of change is readily ascertained. Any conventional 30 technique i8 useful for ascertaining or approximating the slope of such linear data. In one useful technique, starting with any appropriate start point on the curve, e.g., 60 sec., the computing means compares the rate for each time ti of scanning (tl, t2, etc., from the 35 above example) against the rate of the previous time ti 1~ and determines if the rate of change is changing i ~79777 beyond a predetermined limit (for example, + 5%). When the predetermined limit is reached, the rates of change for the times following the time at which the limit is reached, are discarded. The previous rates are then averaged to produce an overall approximation rate.
Conventional subroutines and apparatus, such as microprocessor 100', are used to carry out the approx-imating step described above.
In the aforedescribed process, the values rep-resenting the reflectances and/or reflection densitiesgenerated by the reference liquid are determined either before, during, or after the readings of the test liq-uids. In any event, the reference liquid values are preferably stored in the memory units of computing means 100 or microprocessor 100'.
Most preferably, the values for the reference liquid are determined during calibration. That is, it is a practice to calibrate on a periodic basis, for example, once a week. In this procedure, calibrator liquids having known enzyme activity levels are tested for reflection density or reflectance responses. The procedure of the invention described above is applied to the calibrator liquids, so that each rate of change that is finally ascertained, after subtracting the reference liquid's reflection density responses, is assignable to a known enzyme activity. It is during this procedure that the reference liquid is preferably tested. For example, the calibrator containing the lowest level enzyme activity is useful as the reference liquid. As will be readily appreciated, such a reference liquid is particularly preferred if a change in the sign of the rate is to be avoided when the calibration curve is prepared. Preferably, several replicates of the reference liquid are measured for reflection density responses and an average value taken as the value to be subtracted for each of the plural times.
~7g~777 In preparing a cal~bration curve as a plot of enzyme units per liter (U/L) versus rate, the reference liquid, following subtract~on, produces a zero rate or slope for the enzyme activity chosen as the reference level. The "zero rate" data point thus obtained for the reference liquid iB used together with the known enzyme activity of the reference liquid, as one of the data points from which the calibration curve i8 regressed.
Figs. 3A and 3B illustrate logic flow chart6 that are useful in programming microprocessor 100' to perform the aforedescribed process. From these flow charts a program routine i8 readily determinable u6ing conventional programming techniques. Specifically, sub-scripted variable ti represents the plural read times at which the reflectances and eventually reflection den-sity values are ascertained. The readings are taken over n time units, box 300, Fig. 3A, using m intervals, for example, 5-second intervals. Any start time is useful, it being assumed here that i=l represents the first second after a preliminary, non-reading period is over. The calibration procedure is started, and at each ti, ~he detectable signal (e.g., reflectance) is read for the reference liquid, box 302. The read values are stored in computing means 100 or 100' as values RLi, box 304. Other calibration steps then proceed, whereby a calibration curve of rate of change of reflection density versus enzyme activity is established as internal data within computing means 100 (or 100').
As an additional part of the proces6 of the invention, each test liquid, preferably including each of the calibrator liquids, is examined for its detec-table signal values (e.g., reflectance) at each ti, the same times being used as were used for the reference liquid. Thus, the same preliminary, non-reading period is used, bo~es 310 and 312, Fig. 3B, and the readings taken at the same start time and at the same m inter-vals, for example, 5 seconds, as were used in examining the reference liqu~d. The stored value RLi of the 1 ~9777 reference liquid i6 retrieved for each o~ theæe times, box 313. Before the subtractlon step of the invention is applied, in one embodiment of the invention the values to be modified, e.g., reflectance values, are all converted, box 314, to reflection density values, DR, as described above. If the enzyme iB preferably evaluated as a transmission density, box 315, then the logic branches to a step that converts each DR to DT, as described above. Thereafter, each DR (or DT) so obtained from RLi is subtracted from each corresponding read value of the test liquid, box 316.
The subtraction step is preferably done by the pro-grammed computing means 100 or microprocessor 100'.
When this subtraction is completed for all times ti, optionally the program branches, as shown by the dashed lines, to examiné, box 318, the difference values to ascertain if any value is less than zero. In such a case, since there is no real negative density, it i8 useful to add a constant ~the same constant for all sub-tractions) in an amount such that all difference valuesare non-negative, box 320.
After all times ti have been processed, the density values modified by the subtraction step as ~ust described are optionally stored, box 330, and subse-quently operated on by the subroutine for evaluating theslope.
The foregoing process is repeated for each liquid.
Equivalently, reflectance rather than reflec-tion densities are useful as the values to be appliedduring the linearizing step. In such case, since sub-traction of two logs equals the log of the quotient of the arguments within the two logs, in this embodiment wherein reflectances rather than DR values are used, the subtraction step proceeds as follows: The reflec-tance values of each test liquid are each divided by the respective reflectance values of the reference liquid for the corresponding time ti of the reaction, and the 1 17~377 negative log10 of the quotient is the value to be exam~ned, at each time ti, for slope determination.
The composition of the reference liquid pro-viding the values used in the 6ubtraction step, is not critical. For example, any source of the enzyme i8 use-ful. Any matrix i~ useful, provided it is compatible with the test liquid.
Examples The following examples further illustrate the 10 inventiOn.
Example 1 Multilayered analytical elements for the deter-mination of lactate dehydrogenase were prepared as follows:
A sheet of subbed poly(ethylene terephthalate) was coated with a reagent layer comprising deionized gelatin, 0.43 g/m2 (4.8 mMolar) of NADH, N-[2-hydroxy-1,l-bls(hydroxymethyl)ethyl]taurine buffer (hereinafter TES), octylphenoxy polyethoxy ethanol (a non-ionic surfactant) and a gel hardening agent. Over this wa6 coated a spreading layer compri6ing microcrystall~ne cellulose obtained from FMC Corp. under the trademark "Avicel", poly(vinyl pyrrolidone) obtained from GAF
Corp., the TES buffer noted above, and G.l g/m2 (7.5 mMolar) of ~odium pyruvate.
Onto each element was metered a 10 ~1 drop of calibration liquid comprising human ~erum and a pre-determined level of LDH obtained from rabbit muscle, namely, one of the activity levels 8, 217, 437, 547, 605, 1134, 1685, 2376 and 3071 U/L. These enzyme activities were determined using a6 a reference assay the procedure described in Clin. Chem., Vol. 24, No. 2, pp. 261-266 (1978) modified for the Rotochem Analyzer.
~ 179777 The test elements were placed in an incubator that maintained ~he temperature at 37C. Each of these elements containlng the calibrator was examined every 5 seconds (ti) in an analyzer similar to that shown in Fig. 1. A plurality of reflectance readings were taken at each ti and a mean value of reflectance as-certained. Four replicates of each liquid were used following this procedure, and averaged. For comparative purposes, the DR values for these reflectances were obtained as the negative log10 of the reflectances, and plotted to give markedly non-linear curves of Flg.
4. Although the curves are shown as being continuous, they were in fact the 5 sec. values connected point to point by straight lines. The numbers at the right-hand ends of the curves identify the enzyme activity level.
Next, as the reference values, each of the reflectance values so obtained for the calibrator con-taining 8 U/L of LDH was divided into the reflectance values obtained for the other calibrators at the corresponding times ti, to give modified values every five seconds. The negative logl~ was taken of these modified values and when sequentially connected by straight line~, the resulting DR values formed the curves of Fig. 5. The enzyme levels again are listed at the right-hand ends of the curves. Although the reference having an activity of 8 U/L was not actually plotted, if it were it would appear as the dashed horizontal straight line shown.
It will be readily apparent that, following the fiubtraction step, the reflection density profiles for all enzyme levels were more ~inear than the reflection density profiles of Fig. 4. Of particular interest is Occasionally a data point may not be read at a particular enzyme level. In such a case the value for that time ti is interpolated.
l~97~7 the fact that the ~our levels 217, 437, 547 and 605 U/L
were substantially linear for their entire range.
The linearized curves of Fig. 5 are readily measured for slope, or rate of change, by the techniques noted above. In contrast, any attempt to use the curves of F~g. 4 produces a non-constant and therefore non-unique result.
Because of the superior linearity of the curves of Fig. 5, the dynamic range that is readily detectable 0 i6 expanded to include higher levels of enzyme activ-ity. Such higher levels heretofore were difficult to assay because of the non-linearity of their curves.
Furthermore, in the results of Fig. 5, the analyzer can make its reading with preci6ion anywhere between 15 and 75 sec., regardless of the particular enzyme activity level. However, using the curves of Fig. 4, the curva-ture at 60 sec. for the low levels of activity would prevent the rate from being uniquely determined at the 60 sec. time, for example.
Most preferably, however, this invention allows the analyzer to read as far out as linearity remains in the data, following subtraction of the reference liq-uid's values. For example, the slope can be examined over the time 15 sec. to about 4 minutes, for LDH acti-vity of 605 U/L, without detecting a sufficient depar-ture from linearity as to stop the slope calculation.
However, in the case of LDH at 3071 U/L, the slope can be measured only out to about 90 sec. before curvature causes the slope to change at a rate exceeding pre-assigned limits. Thi6 ability to read the slope for aslong as linearity iB preBent permits the rate to be read over a longer time, particularly at lower enzyme activity levels. This is advantageous becau6e the imprecision due, e.g., to instrument noi6e, is thereby reduced.
l 1~9777 Example 2 - Use in Assaying for LD-4 Isoenzyme The procedure of Example 1 was repeated, except that the calibrator liquids metered onto the test ele-ments each contained various predetermined levels of LD-4 isoenzyme obtained from human sources, having the activities 11, 110, 386, 788, 1458 and 2223 U/L. The response values representing the corresponding reflec-tion densities appear in Fig. 6 as markedly non-linear curves. To improve these curves, each of the 5-second interval reflectance values of the curve for the 11 U/L
activity were divided, as reference values, into the reflectance values for the corresponding 5-second reading of the other curves. The negative log10 of these modified values produced modified DR values, which appear in Fig. 7. It will be readily apparent that, following the subtraction step, the reflection density profiles for all enzyme levels were more linear, which is in marked contrast to the profiles of Fig. 6.
The difference values for the 110 and 386 U/L activity are substantially linear for the entire rdnge.
Example 3 - Use in Assaylng for AST
The procedure of Example 1 was repeated, except that the test elements were constructed to assay for aspartate aminotransferase (AST) as follows:
A polyethylene terephthalate film was coated with a reagent layer comprising deionized gelatin, surfactant, gel hardening agent, tris(hydroxymethyl)-aminomethane (tris) buffer, NADH (0.32 g/m2 or 3.6 mMolar), pyridoxal-5-phosphoric acid (PLP) (0.16 g/m~), malate dehydrogenase (MDH) and lactate dehydrogenase (LDH); a subbing layer comprising poly-(N-isopropylacrylamide); and a spreading layer com-prising barium sulfate, cellulose acetate, the non-ionic ~urfactant of Example 1, a polyurethane resin elastomer obtained from B.F. Goodrich, sodium ~-ketoglutardte and sodium asparate.
~ 179777 The calibrator liquids comprised a bovine serum albumin matrix in which the AST levele were ad~usted using porcine heart AST, so that samples had one of the following predetermined levels of AST activity: 25, 641~ 1246, 2316, and 4777 U/L. Fig. 8 is a plot of the DR response curves for the reflectances originally detected. Thereafter, each of the 5-second reflectance values of the curve for the 25 U/L activity, the ref-erence liquid, was divided into the reflectance values obtained at the corresponding time for the other curves. When converted into reflection densities the resulting modified values appear in Fig. 9. Again, marked improvement in linearity was noted.
Example 4 - Use of a High Level Enzyme Activity Ref-erence Liquid The procedure of Example 1 was repeated, exceptthat a calibrator liquid having 1292 U/L activity of LDH
was selected as the reference liquid. Before modifying the reflectance values using the reflectances obtained for the reference liquid, the reflection density response curves were as shown in Fig. 10. After division of reflectances ~equal to subtraction of reflection densities), the corresponding reflection density curves appeared as shown in Fig. 11. Even those levels above 1292 U/L demonstrated, after subtraction, a longer time span of linearity than was available before subtraction. E.g., the highest level at 3840 U/L was generally linear after subtraction out to 1.5 minutes, whereas before it was generally linear only out to 1.0 minute.
In carrying out the aforedescribed analysis of the slope of the curves of Fig. 11, the tail portions labeled "B"', for levels 2425, 3320 and 3840 U/L, would l~797 of course not be used, as they represent the reactant depletion region.
Example 5 - Use of a Non-Human Reference Liquid Essentially the procedure of Example 1 wa6 S repeated except that the test element formulation wa6 that of Example 2, and the calibrators were prepared from a bovine serum pool that was 6tripped of LDH and spiked w~th LDH obtained from porcine heart to achieve the desired levels. Fig. 12 is the graph of the response curves of the calibrators before the reference liquid's values were Rubtracted, point-by-point. Fig.
13 is the corresponding graph produced following sub-traction of the values for the 16 U/L activity, selected to be the reference liquid. As before, the linearity was markedly improved by the subtraction step.
Example 6 The procedure of Example 1 was repeated, except that the LDH samples tested were actual patient ~erum samples, rather than serum-based calibrators. A marked improvement in linearity was observed in the data after subtraction of the reference liquid' 6 values, point-by-point, just as in the case of Example 1.
Example 7 The procedure of Example 4 was repeated, except that the reflection density DR values for the enzyme activity 341 U/L, Fig. 10, were divided by the reflection density DR values for the liquid having an activity level of 45 U/L, as the reference liquid for this example. Thus, the arithmetic operation was divis-ion, whereas in Examples 1-6 previously described, the operation was subtraction. The resulting plot of the modified DR was more linear than the original plot for 341, Fig. 10.
The invention has been described in detail with particular reference to preferred embodiments thereof, ~ ~ ~9~7~
but lt will be understood that variations and modifi-cations can be effected within the spirit and scope of the invention.