JP3496390B2 - Data processing method for glycated hemoglobin measuring device using liquid chromatography - Google Patents

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a data processing method of a glycated hemoglobin measuring apparatus using liquid chromatography.

[0002]

2. Description of the Related Art Conventionally, the concentration of glycated hemoglobin in blood has been measured by using a liquid chromatography apparatus as shown in FIG. First, the glycated hemoglobin component in the blood sample 2 collected by the sampling unit 1 is once adsorbed to the separation system 3 of liquid chromatography including a column packed with a cation exchange resin, separated from impurities, and then eluted. Elute out of the separation system by liquid. The eluted component is detected by the detector 4 as a change in absorbance.

FIG. 2 shows a chromatogram of a standard sample as an example of the chromatogram (change in absorbance depending on elution component depending on elution component) obtained by the conventional measuring method as described above. As shown in FIG. 2, in the measured chromatogram, there are some peaks corresponding to the constituent components of glycated hemoglobin in blood. For example, A
The 1c peak is a peak corresponding to the component in which glucose is added to the N-terminal of the beta chain of hemoglobin, and the A0 peak is a peak corresponding to the component of non-glycated hemoglobin.

The chromatogram data is A / in FIG.
After being converted into a digital signal by the D converter 5, the CPU
The data is processed according to 6.

Finally, the data processing unit 6 including the CPU obtains the peak area of each peak by the peak detection processing method as shown in FIG. 3 and provides information for various diagnoses of diabetes.

In order to obtain the peak area of each peak, conventionally, as shown in FIG. 3, a baseline for area calculation is first set. Chromatographic data The differential value of each point, that is, the slope of the tangent line is calculated, and the point at which the differential value becomes larger than the predetermined threshold value (positive value) is set as the baseline start point, and the differential value is also set in advance. The point that becomes smaller than the threshold value (negative value) is the baseline end point. The straight line connecting the baseline start point and the baseline end point is used as the peak baseline, and the area of the region surrounded by the chromatographic data curve and the baseline is used as the peak area value of this peak.

Further, as shown in FIG. 4, when two or more peaks are overlapped with each other, a straight line having a predetermined slope (for example, the lowest point between the peaks and the baseline start point) (eg, The area is calculated by changing the baseline depending on whether it is located above or below the straight line indicated by the broken line in the figure). That is, when the lowest point is below the above straight line (when facing left)
Uses the lowest point as the baseline end point of the first peak (peak on the left side) and the baseline start point of the second peak (peak on the right side), and the base end point of the second peak is the same as in the case of a single peak. To decide. On the other hand, when the lowest point is located above the straight line (when facing right), the baseline is the first peak (left peak).
And the second peak (peak on the right side) are common, the baseline end point is determined in the same manner as in the case of a single peak in the second peak, and the straight line drawn from the lowest point perpendicular to the baseline is the first peak. It is defined as the boundary line of the second peak (hereinafter, in the present specification, the data processing method when the lowest point is located above the straight line is referred to as trough cutting peak processing).

[0008]

There are various problems in the conventional data processing method as described above. Generally, there are individual differences in the hemoglobin concentration in blood, and the absolute concentrations of each component differ from individual to individual, but in the diagnosis of diabetes by a glycated hemoglobin measuring device using liquid chromatography, A1c, etc. for the total hemoglobin concentration in blood, etc. Since the diagnosis is performed by using the concentration ratio (peak area ratio) of the components, the same concentration ratio must be given even when the component ratios are the same and the concentrations are different.

However, in the conventional baseline setting work for data processing, the same is true for samples of different concentrations having different thresholds for determining the baseline start point and the baseline end point from the differential value. Therefore, a deviation from the peak area value that should be originally obtained may occur.

For example, an ideal Gaussian curve peak Y =
Assuming that C · exp (−0.001 · X · X), the peak A with a peak height of 1 (that is, C = 1) and the peak B with a peak height of 10 (that is, C = 10) are identical. If the baseline start point is determined with a threshold value of 0.001 and the baseline end point is determined with the same threshold value of -0.001,
As shown in the table below, a deviation from the theoretical area ratio that should be originally obtained occurs.

[0011]

[Table 1]

Further, as shown in FIG. 2 described above,
In the chromatogram of glycated hemoglobin,
Since multiple peaks are measured overlapping each other, a predetermined straight line for determining whether or not to make a valley peak for many peaks was the same even for samples of different concentrations, so it is originally required. The deviation from the peak area value that should have occurred sometimes occurred.

The above-mentioned deviation from the true area of the data processing result occurs even with an ideal Gaussian curve, but since the actual measurement data contains a noise component (peak). However, when the glycated hemoglobin concentration is low, the S / N ratio of the signal is low, and the possibility of deviation is high.

Further, in the actual measurement of glycated hemoglobin, for example, as shown in FIG. 5, the last part of the peak derived from a component called A0, which occupies 90% of the total peak area appearing at the end of the chromatogram, is Even with the same sample, the solid line or the broken line may appear due to deterioration of the column or slight changes in the concentration of the eluent. When valley truncation peak processing including A0 peak is performed as shown in FIG. 5, even if the peak shape other than A0 does not change, the baseline end point changes greatly due to the change in the last part of A0 peak, The baseline moves. As a result, A0 sharing the baseline
Other peaks other than the above are affected, and the peak shape does not change. Therefore, there arises a problem that the area value of the peak should change without changing the area value.

Moreover, since the area of the A0 peak is usually about 10 to 50 times larger than the other peaks, the area changed by the baseline fluctuation is negligible in the case of the A0 peak, but it is a trace component. It is not possible to ignore changes in the peak area values that are important for diabetes diagnosis, such as certain A1c.

[0016]

The present invention has been made in view of the above problems, and its object is to provide an accurate and reproducible A1c area ratio for a sample having a wide glycated hemoglobin concentration distribution, etc. It is to provide a data processing method for giving the component area ratio of

That is, the present invention is a liquid chromatography in which a blood sample is subjected to liquid chromatography to obtain an absorbance chromatograph, and the ratio of the total area of the elution peak in the chromatogram to the peak area of glycated hemoglobin as a target component is obtained. A method for processing data of a glycated hemoglobin measuring apparatus using, wherein (A) a step of standardizing a chromatogram to perform peak detection processing, and (B) a step of performing a smoothing processing with weighting according to the maximum absorbance value of the chromatogram. , (C) Step of setting the baseline of peak area calculation by the absorbance at the baseline start of the first peak, and making the line of equal absorbance as the baseline in all peak area calculations, (D) Break point between each peak Is defined as the time at which the absorbance is the lowest between the peaks, and the peaks are Dividing and calculating each peak area, and (E) determining the end point of the final peak by the time from sample injection, dividing the final peak by isochronous lines, and calculating the peak area. A method characterized by including steps. The present invention will be described in detail below.

In the standardization step of the present invention, the difference between the minimum value and the maximum value of the chromatogram data obtained by liquid chromatography is X, and the entire chromatogram has a minimum value of 0 and a maximum value of A. Is standardized according to the following formula.

[0019]

[Equation 1]

A chromatogram obtained by a glycated hemoglobin measuring apparatus using liquid chromatography usually occupies about 90% of the total area as shown in FIG.
A1c occupies several% of the total area before the component called 0
And so on. Therefore, usually, X in the equation 1 is almost equal to the height of A0 peak and is almost proportional to the total area value. Therefore, this normalization almost completely eliminates the error caused in the area value due to the concentration of the whole glycated hemoglobin in the blood, and the ratio (%) of these minute peaks to the total area used in the diagnosis of diabetes is accurately and reproducibly. It becomes possible to ask.

After the standardization is performed, the standardized data is subjected to a smoothing process weighted by the size of X. The smoothing process is performed by using a low-pass filter for the purpose of noise reduction. The smaller the X, the lower the cut-off frequency of the low-pass filter, or the greater the cut-off attenuation rate, to reduce the S / S of the original data.
The lower N is, the more the S / N is improved, and the S / N depends on the X value.
It is intended to reduce the difference in the ratio.

The weighting method may be determined according to the type of low-pass filter and the variation in the value of X, but preferably one step, that is, when weighting is not required, more preferably two or more steps Should be provided.

Generally, a low-pass filter is used to improve the S / N ratio of data by attenuating a frequency component higher than the cutoff frequency in the data and to perform smoothing. That is, in general liquid chromatographic analysis, the frequency component caused by noise is on the higher frequency side than the frequency component caused by the chromatogram peak, so the frequency caused by this chromatogram peak as the cutoff frequency of the low-pass filter. By selecting the frequency where most of the components are included, it is possible to attenuate only the noise component without affecting the peak of the chromatogram. Also, since the attenuation rate for high frequency components depends on the characteristics of the filter,
The noise removal effect can be enhanced by using a filter having a high attenuation rate.

With respect to the standardized data as described above, since the raw data having the same noise width is multiplied by the coefficient of the value of X, the noise becomes smaller as the value of X becomes larger, and the noise of X becomes larger. The smaller the value, the larger the value. Therefore, the same S / N for standardized data of different concentration samples
In order to calculate the area by the ratio, a filter having a low attenuation rate when the value of X is large and the noise is small, and a low-pass filter having a high attenuation rate when the value of X is small and the noise is large depends on the value of X. Perform weighting. This weighting step may be determined by the variation in the X value. When the X value range is narrow, it is not necessary to perform one step, but when the X value range is wide, there are two or more steps. It is preferable to use a low-pass filter having a plurality of attenuation factors.

A first differential value is calculated for the smoothed data of the standardized data, and a point exceeding a predetermined threshold value is set as a baseline start point of the first peak, and the baseline is from this start point. The absorbance is determined horizontally, ie at the baseline start point, and the line of equal absorbance is taken as the baseline for all peak area calculations. This threshold is usually A x 0.0003
It may be about (1 / min), but A × 10 ^{−5 to} A × 10
^A baseline of ^-2 (1 / min) can be optimized, which is preferable.

By determining the baseline start point from the standardized data, the baseline start point can be accurately determined regardless of the value of X. Furthermore, by setting the baselines to have the same absorbance (horizontally drawn), even if the shape of the last peak of A0 changes, the baseline does not change.

Between the peaks after the second peak, the time at which the absorbance is minimum is determined between the peaks, and each is divided by a line of equal time. Then, the area of each peak surrounded by this isochronous line and the baseline is calculated. The obtained area value is converted into the unit before normalization according to the following equation using the parameters X and A used in normalization. However, this step is necessary only for obtaining the area value, and is unnecessary if only the area ratio of each peak needs to be obtained.

[0028]

[Equation 2]

In the chromatogram obtained by measuring glycated hemoglobin, the A0 peak appears last. As described above, the final point of this peak changes due to subtle column deterioration and denaturation of the eluent. Therefore, in the present invention, the end point of the final peak is set to a fixed time after the glycated hemoglobin component starts to elute from the column, and the final peak is divided by a line of equal time to calculate the peak area. The certain period of time may be determined so that a standard sample is measured under the liquid chromatography conditions for measuring glycated hemoglobin, and the elution of the A0 peak in the obtained chromatogram is completely completed. , The type of eluent for eluting the glycated hemoglobin component from the column,
Furthermore, it is appropriately determined according to the speed of the eluent delivery. For example, when a cation exchange resin is packed in a column having a volume of 0.6 ml and a buffer solution such as succinic acid is delivered at a rate of about 1.5 ml / min, the delivery of the eluent of the target component to the column is started. It can be exemplified that it is set to about 2.2 minutes.

As described above, (A) standardization step, (B) smoothing step, (C) baseline setting step, (D)
Although the steps for setting the boundary line of each peak and (E) dividing the final peak (A0 peak) are included, these steps do not necessarily have to be performed in that order, and (A) and (B ) Step,
There is no special limitation as long as it is performed prior to the steps (C) and (D).

[0031]

EFFECTS OF THE INVENTION According to the present invention, since it is possible to always perform chromatogram data processing on samples having different concentrations under the same conditions, it is possible to accurately and reproduce human blood samples having a wide glycated hemoglobin concentration distribution. It is possible to calculate the peak area with certainty. As a result, it is possible to provide accurate and reproducible results for diagnostic parameters using the area ratio such as A1c used for diabetes diagnosis.

Further, even if there is a slight column deterioration or denaturation of the eluent, it is possible to minimize the influence on the area value of a minute peak such as A1c which is important in the diagnosis of diabetes. It is possible to provide the area ratio of the A1c peak used for diabetes diagnosis as accurate and reproducible.

[0033]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described in more detail with reference to the drawings, but the present invention is not limited to the invention described in these drawings.

FIG. 1 is a diagram showing the structure of a glycated hemoglobin measuring apparatus. The present invention relates to a data processing method in the CPU 6 (data processing unit) in FIG. 1, and a conventional device can be used. As shown in FIG. 1, the glycated hemoglobin component in the blood sample 2 collected in the sampling unit 1 was once adsorbed to the separation system 3 of the liquid chromatography composed of a column packed with a cation exchange resin, and was detected as a contaminant. After the separation, the eluent elutes outside the separation system. The eluted component is detected by the detector 4 as a change in absorbance. Then, the obtained chromatogram data is converted into a digital signal by the A / D converter 5, and then processed by the CPU 6.

FIG. 6 shows a flowchart in the CPU 6 (data processing section) for carrying out the data processing method of the present invention. In the CPU 6 (data processing unit), first,
The maximum value and the minimum value in the data are obtained, X, that is, the difference between the maximum value and the minimum value is calculated, and then the standardization is performed using the equation 1 described above. The standardization parameter A is, for example, A = 1.
00 etc., but if the area calculation is performed by integer arithmetic, A
= 1,000,000 or the like, which is changed according to the calculation capability of the CPU.

Next, using a digital filter (low-pass filter), a cutoff frequency of about 0.014 [Hz]
Noise removal is performed with and smoothing processing is performed. The cutoff frequency may be changed according to the frequency component of the peak of interest in the chromatogram, for example, it may be reduced to about half when the time width of the peak of interest is doubled.

Next, a point where the differential value of the data exceeds a predetermined value is set as a baseline start point, and an absorbance line equal to the absorbance value at this point is drawn as the baseline of all peaks. As is clear with reference to FIG. 2 and the like, drawing a line of equal absorbance is nothing but drawing a horizontal line. Although the threshold value in this step can be arbitrarily determined, it is usually A × 0.000.
It may be about 3 (1 / min), and A × 10 ^{−5 to} A × 10
^The baseline can be optimized by setting ^-2 (1 / min).

Next, the time at which the absorbance becomes the minimum between the peaks is determined, and the peaks are divided by isochronous lines. As is clear from FIG. 2 and the like, drawing lines of equal time is nothing but drawing vertical lines. As a result, the previously drawn baseline and the boundary of each peak drawn in this step intersect vertically.

Although not shown in FIG. 6, the final peak (A
The end point of 0 peak) is determined by the time from the start of elution of the glycated hemoglobin component, and the final peak is divided by isochronous lines.

The area of each of the peaks thus divided is calculated, and the ratio is calculated. When calculating the area value as necessary, each parameter may be substituted into the above-mentioned formula 2.

[Brief description of drawings]

FIG. 1 shows the configuration of a fully automatic glycated hemoglobin measuring device.

FIG. 2 is a diagram showing a chromatogram obtained when an appropriate glycated hemoglobin measurement is performed on a standard sample.

FIG. 3 is a diagram showing a conventional peak detection data processing method.

FIG. 4 is a diagram showing a conventional peak detection data processing method when two or more peaks overlap.

FIG. 5 is a diagram for explaining a conventional problem that occurs when the shape of the A0 peak changes.

FIG. 6 is a diagram showing a flowchart of a data processing method of the present invention.

[Explanation of symbols]

1 Sampling part 2 Blood sample 3 Separation system for liquid chromatography consisting of a column packed with cation exchange resin 4 Detector 5 A / D converter 6 CPU (data processing part)

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁷ Identification code FI G01N 30/86 G01N 30/86 L (56) References JP-A-5-5730 (JP, A) JP-A-8-233795 ( JP, A) JP 7-120449 (JP, A) JP 9-127091 (JP, A) JP 2000-321261 (JP, A) JP 5-60742 (JP, A) JP 6-324029 (JP, A) JP-A-2-221859 (JP, A) JP-A-9-251016 (JP, A) JP-A-3-255360 (JP, A) JP-A-4-77500 (JP, A) A) JP-A-7-120447 (JP, A) JP-A-4-156952 (JP, A) JP-A-4-164250 (JP, A) JP-A-6-34634 (JP, A) JP-A-9 −178719 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G01N 30 / 00-30 / 96

Claims (4)

(57) [Claims] 1. A liquid chromatograph is used in which a blood sample is subjected to liquid chromatography to obtain an absorbance chromatograph, and the ratio of the total area of the elution peak in the chromatogram to the peak area of glycated hemoglobin as a target component is used. A data processing method for a glycated hemoglobin measuring apparatus, comprising the steps of (A) normalizing a chromatogram to perform peak detection processing, (B) performing a smoothing processing with weighting according to the maximum absorbance value of the chromatogram, (C) ) A step of setting the baseline for peak area calculation by the absorbance at the baseline start of the first peak, and making the line of equal absorbance as the baseline for all peak area calculations, (D) Break points between peaks Is determined by the time at which the absorbance is the lowest, and the peak is divided by the line of equal time. The step of calculating these peak areas, and (E) the end point of the final peak (A0 peak) is determined by the time from the start of the elution of the glycated hemoglobin component, and the final peak is divided by isochronous lines to calculate the peak area. And a step of performing the steps of. 2. The data processing method according to claim 1, wherein the liquid chromatography uses a column packed with a cation exchanger. 3. The data processing method according to claim 2, wherein the column volume is about 0.6 ml, and the liquid feed to the column is liquid chromatography of 1.5 ml / min. 4. A buffer solution for adsorbing glycated hemoglobin in a blood sample on a column packed with a cation exchange resin is succinic acid of 50 mM or less, and a buffer solution for eluting the glycated hemoglobin adsorbed on the column. Is 0.1
The data processing method according to claim 2, wherein the succinic acid is 0.5M.