CN103792205B - The sensitive quick nondestructive analysis of the high flux near-infrared of tablet impurity and tensile strength - Google Patents

Abstract

The invention relates to a high-throughput near-infrared sensitive, rapid and non-destructive analysis method for tablet impurities and tensile strength, belonging to the field of pharmaceutical preparation analysis. Include the following steps: prepare or collect tablet sample; Collect the near-infrared diffuse reflectance spectrum (NIR-DRS) of known tablet sample with Fourier transform near-infrared spectrometer; Measure tablet with high performance liquid chromatography (HPLC) Reference value of impurity content; measurement and calculation of reference value of tensile strength with vernier caliper and tablet hardness tester; preprocessing of spectra; selection of optimal modeling wavenumber range and removal of singular values; establishment of impurities based on NIR‑DRS respectively Calibrate the model with tensile strength and evaluate the model performance; collect unknown tablet sample NIR‑DRS; perform the same pretreatment on unknown tablet sample NIR‑DRS as known sample NIR‑DRS; use the built model to predict unknown tablet The impurity content and tensile strength of the agent sample. The method has high sensitivity, no need for sample pretreatment, fast and non-destructive analysis, and accurate results.

Description

High-throughput near-infrared sensitive and rapid non-destructive analysis of tablet impurities and tensile strength

technical field

The invention relates to a method for rapidly detecting impurities and tensile strength in pharmaceutical tablets, specifically a method for rapidly and nondestructively detecting impurities and tensile strength in tablets using Fourier transform near-infrared spectroscopy combined with chemometrics technology The method belongs to the field of pharmaceutical preparation analysis.

Background technique

Tablet refers to a sheet-like preparation formed by uniform mixing of drugs and pharmaceutical excipients. Tablets have the advantages of convenient taking, good chemical stability, accurate dosage, and convenient carrying. At present, among the preparations recorded in the pharmacopoeias of various countries in the world, tablets are the most.

During the production and storage of tablets, some impurities are often introduced, affecting the stability and curative effect of tablets, and even endangering human health. Therefore, it is necessary to carry out the impurity inspection of tablets to ensure their safety and effectiveness, and also provide a basis for the supervision and management of drug quality in the production and distribution process. In the pharmacopoeias and documents of various countries, the methods for determining the impurity content of tablets are mostly high-performance liquid chromatography (HPLC), gas chromatography (GC), liquid chromatography-mass spectrometry (LC-MS), gas chromatography-mass spectrometry (GC-MS )Wait. The above-mentioned analysis method involves the weighing of tablet weight, the preparation of the test sample solution, and the analysis for a long time. The efficiency is low, and the sample is damaged, so it cannot realize automatic online quality control.

Tensile strength, an important tablet physical parameter, can reflect the disintegration time limit and dissolution rate of the tablet, etc. The calculation formula of the tablet tensile strength (radial tensile strength, RTS) is:

Wherein, F is the hardness of the tablet, D is the diameter, and t is the thickness. Vernier calipers are often used to measure thickness and diameter, and hardness testers are used to measure hardness. The measurement process involves the measurement of tablet diameter and thickness, the determination of tablet hardness, and the calculation of tensile strength. The operation is cumbersome and destroys the sample, so it is not suitable for the rapid analysis of a large number of samples and the online control of the process.

Near-infrared spectroscopy characterizes frequency-doubled and combined-frequency absorptions of hydrogen-containing groups in molecular vibrational transitions. Near-infrared spectroscopy is a non-destructive detection and analysis technology with the advantages of high-throughput, rapidity, simplicity, and automation. It can analyze various chemical and physical information in samples at the same time. has a wide range of applications.

Currently, there is no literature on high-throughput sensitive near-infrared spectroscopy analysis of impurities and tensile strength in tablets.

Contents of the invention

The object of the present invention is to provide a near-infrared spectrum analysis method for tablet impurities and tensile strength that is sensitive, fast in analysis speed, easy to operate, pollution-free and automatic. It mainly includes the following steps:

1. Prepare or collect tablet samples;

2. Collect the near-infrared diffuse reflectance spectrum (NIR-DRS) of the tablet sample;

3. Determine the content reference value of impurities in the tablet sample by HPLC;

4. Measure the diameter, thickness and hardness of the tablet sample with a vernier caliper and a tablet hardness tester, and calculate the tensile strength reference value;

5. Preprocess the original NIR-DRS;

6. Select the optimal modeling wavenumber range and eliminate singular values;

7. Select the optimal number of factors to establish the impurity content and tensile strength correction model based on the NIR-DRS of the tablet sample, and evaluate the model performance;

8. Acquisition of NIR-DRS for Unknown Samples

9. Perform corresponding pretreatment on NIR-DRS of unknown samples;

10. Apply the built model to predict the impurity content and tensile strength of unknown samples.

Among them, the NIR-DRS of the tablet is usually collected by a Fourier transform near-infrared spectrometer. The sampling device can use the diffuse transmission sampling accessory of the tablet. The signal collection software includes but not limited to: Result3.0, and the data processing software includes but not limited to: TQAnalyst8 .0. Spectrum acquisition modes include, but are not limited to: NIR-DRS. The resolution of NIR-DRS measurement parameters includes but not limited to: 8cm ^-1 , the number of scans includes but not limited to: 64, and the spectral range includes but not limited to: 10000 ~ 4000cm ^-1 .

Preprocessing methods for raw spectra and data include, but are not limited to: unprocessed, multivariate scatter correction or standard normal transformation, first or second derivative, Norris smoothing or Savitzky-Golay smoothing, mean centering, scaling, etc. One or more of the above methods are used in combination to achieve the best model prediction performance.

The spectral wavenumber range used for modeling can be automatically screened by the modeling software, and the automatically screened range can be manually optimized according to the near-infrared characteristic absorption of the analyte, so that the calibration model has low prediction error and high correlation coefficient.

Singular values can be automatically screened and eliminated by modeling software, and manually screened and eliminated based on spectral data, reference values or predicted effects.

Based on NIR-DRS, methods for establishing a calibration model of impurity content or tensile strength in tablet samples include but are not limited to: Partial Least Squares (PLS). Determine the optimal number of principal factors for the PLS model by cross-validation, and use the correction set root mean square error (RMSEC), cross-validation root mean square error (RMSECV), prediction set root mean square error (RMSEP), correction set correlation coefficient (R _c ) and prediction set correlation coefficient (R _p ) to evaluate the performance of the calibration model.

The parameter values of NIR-DRS acquisition of unknown samples are the same as the acquisition method of spectrum of known samples in the calibration model. Based on the NIR-DRS of unknown samples, the model can be used to predict the impurity content and tensile strength of unknown samples in a high-throughput, sensitive and rapid manner.

This method is suitable for rapid analysis of one or more impurities and tensile strength in tablets, without complicated sample pretreatment, without destroying the sample, and without pollution.

Description of drawings

Figure 1 Structural formulas of reduced glutathione (GSH) and oxidized glutathione (GSSG)

Fig. 2 The original Fourier transform NIR-DRS of 330 samples of reduced glutathione flakes.

Figure 3A and B are the relationship diagram between the number of principal factors of the oxidized glutathione PLS model and the RMSECV, and the linear correlation diagram between the reference value of the prediction set and the predicted value.

Figure 4A and B are the relationship diagrams between the principal factors of the tensile strength PLS model and RMSECV, and the linear correlation diagrams between the reference value of the prediction set and the predicted value.

detailed description

The method of the present invention is applied to the rapid detection of oxidized glutathione and tensile strength in reduced glutathione tablets, and the examples will be described below with reference to the accompanying drawings.

Example

1. Tablet sample preparation and its NIR-DRS collection.

Reduced glutathione tablet samples (100mg/tablet) were prepared by entrusting manufacturers to change the dosage of the main drug according to the registered prescription process.

Instrument: Fourier transform near-infrared spectrometer, the sampling device is a tablet diffuse transmission sampling accessory, the signal acquisition software is Result3.0, and the data processing software is TQ Analyst8.0.

Scanning conditions: Use the integrating sphere tablet diffuse transmission sampling accessory to scan both the front and back of the tablet. Scan the background before scanning the sample.

Measurement conditions: the resolution of NIR-DRS measurement parameters is 8cm ^-1 , the number of scans is 64, and the spectral range is 10000-4000cm ^-1 .

2. Determine the content of oxidized glutathione impurity in the tablet sample by HPLC, as a reference value.

Chromatographic conditions refer to the State Food and Drug Administration standard YBH1824-2004, measure the impurity peak area in the tablet samples and calculate the content of oxidized glutathione in 186 samples by the external standard method.

3. Determine the tensile strength of the tablets.

Measure the thickness and diameter of the sample with a vernier caliper, measure the hardness of the sample with a tablet hardness tester, and calculate the tensile strength of 315 reduced glutathione tablet samples according to the formula.

4. Determination of the content of the main impurity oxidized glutathione in reduced glutathione tablets by partial least squares (PLS).

In order to achieve the purpose of accurate prediction, the calibration set and the prediction set are carefully screened. In the analysis of oxidized glutathione, the spectra of the calibration set and the prediction set are screened at a ratio of 2:1, and the prediction set The reference value of oxidized glutathione content is evenly distributed in the range of the reference value of the calibration set. Automatically check and remove abnormal spectra with TQ Analyst8.0. Finally, it is obtained that the reference value (15.85-22.18mg/g) of the oxidized glutathione content of the 62 prediction set samples is evenly distributed within the range of the reference value (15.81-22.20mg/g) of the 124 calibration set samples; The spectral preprocessing method of the type glutathione content correction model is mean centering.

The modeling wavenumber range is automatically screened by TQ Analyst8.0, and then the modeling wavenumber range is manually optimized according to the root mean square error (RMSEP) value of the prediction set to obtain the best modeling effect. The spectral range of the oxidized glutathione calibration model (8633.39～4133.15cm ^-1 ) includes the frequency-doubled absorption of the main functional groups such as carboxyl, primary amino and amide groups of oxidized glutathione: the dichotomy of C=O in the carboxyl First order frequency peak (5260cm ^-1 ), first order frequency octave of OH in carboxyl group (6920cm ^-1 ), first order frequency octave of NH in primary amino group (6600cm ^-1 ), first order frequency octave of NH in amide group (6803 ～6711cm ^-1 ), and the first order octave of CH (5882～5555cm ^-1 ).

The cross-validation method was used to determine the optimal number of principal factors of the PLS model, and the number of principal factors used was 12, which corresponded to the minimum RMSECV.

The performance of the built model is evaluated by the following parameters: root mean square error of correction set (RMSEC), root mean square error of cross-validation (RMSECV) and root mean square error of prediction set (RMSEP), correlation coefficient of correction set (R _c ), prediction Set correlation coefficient (R _p ).

The performance indicators of the established calibration model are shown in Table 1.

Table 1. Various performance indicators of the oxidized glutathione content correction model and the tensile strength correction model

As shown in Table 1, the RMSEC, RMSECV and RMSEP of the model are 0.575, 0.729 and 0.607 mg/g, R _c and R _p are 0.9173 and 0.9182, respectively. The reference value of the prediction set of oxidized glutathione and the prediction value showed a good linear relationship, and the regression equation of the prediction set was y=0.8614x+2.6982. The error of the built model is small and the correlation coefficient is high, indicating that although there is a high content of the main drug of reduced glutathione with a similar structure in the sample, the low content of oxidized glutathione can still be accurately measured by near-infrared spectroscopy.

5. Determination of tensile strength of reduced glutathione tablets by partial least squares (PLS).

In the analysis of tensile strength, the spectra of the calibration set and the prediction set were screened at a ratio of 2:1, and the reference value of the tensile strength of the prediction set was evenly distributed within the range of the reference value of the calibration set. Automatically check and remove abnormal spectra with TQ Analyst8.0. Finally, it is obtained that the tensile strength reference values (298.65-1119.61kPa) of the 105 prediction set samples are evenly distributed within the range of the tensile strength reference values (270.76-1145.14kPa) of the 210 calibration set samples.

The spectral preprocessing method of the tensile strength correction model is 7 points 3 times Savitzky-Golay smoothing, mean centering and calibration.

The modeling wavenumber range is automatically screened by TQ Analyst8.0, and then the modeling wavenumber range is manually optimized according to the root mean square error (RMSEP) value of the prediction set to obtain the best modeling effect. The spectral range of the tensile strength calibration model is (8974.47～4039.12cm ^-1 ).

The cross-validation method is used to determine the optimal number of principal factors of the PLS model, and the number of principal factors used is 9, which corresponds to the number of principal factors when the RMSECV is the smallest.

The performance indicators of the established calibration model are shown in Table 1.

As shown in Table 1, the RMSEC, RMSECV, and RMSEP of the model are 58.2, 61.3, and 69.0 kPa, respectively, and R _c and R _p are 0.9393 and 0.9151, respectively. The reference value of the tensile strength prediction set has a good linear relationship with the predicted value, and the regression equation of the prediction set is y=0.8211x+113.3224. The error of the built model is small and the correlation coefficient is high, indicating that the physical parameters of the tablet, the tensile strength, can be accurately measured by near-infrared spectroscopy.

The present invention proposes a method for quickly and non-destructively detecting tablet impurities and tensile strength by NIR spectroscopy. The research results show that by establishing a PLS model, the NIR spectroscopy analysis method can detect the oxidized glutathione, the main impurity in the reduced glutathione tablet. Glutathione content and tensile strength are accurately detected. Compared with traditional methods, this method has high selectivity, no need for sample pretreatment, rapid analysis and accurate results. It opens up a new direction for high-throughput, sensitive, rapid and non-destructive analysis of impurities and tensile strength in tablets.

Claims (5)

1. in reduced glutathione piece oxidation of impurities type glutathione and tensile strength near-infrared quick nondestructive analysis method, It is characterized in that adopting following steps：

(1) prepare or collect tablet samples；The tablet diffusing transmission sampling annex collection of application Fourier Transform Near Infrared instrument The double-edged near-infrared diffusing reflection spectrum of tablet samples is NIR-DRS；Acquisition parameter be scanning times 64, resolution ratio 8cm^-1, 10000～4000cm of sweep limits^-1；

(2) with high performance liquid chromatography be HPLC determine tablet samples in oxidation of impurities type glutathione content reference value；Right Original NIR-DRS carries out average centralization pretreatment；Select optimum modeling 8633.39～4133.15cm of wave-number range^-1, reject Singular value；Optimal main cause subnumber 12 is selected to set up the impurity content PLS based on tablet correcting sample NIR-DRS Calibration model, is verified using tablet prediction sample, and is that RMSEC, cross validation are square using calibration set root-mean-square error Root error is RMSECV, forecast set root-mean-square error i.e. RMSEP, calibration set coefficient correlation i.e. R_cAnd forecast set coefficient correlation is R_p Model performance is evaluated；

(3) determined with slide measure and tablet hardness tester and calculate the reference value of tablet samples tensile strength；To original NIR-DRS carries out that 7: 3 Savitzky-Golay are smooth, the pretreatment of average Centering and scaling；Select optimum modeling wave number model Enclose 8974.47～4039.12cm^-1, abnormal value elimination；Optimal main cause subnumber 9 is selected to set up based on tablet correcting sample NIR-DRS Tensile strength PLS calibration model, using tablet prediction sample verified, and using RMSEC, RMSECV, RMSEP、R_cAnd R_pModel performance is evaluated；

(4) NIR-DRS of unknown sample is gathered；The NIR-DRS of unknown sample is pre-processed accordingly；Application institute established model The impurity content of prediction unknown sample and tensile strength.

2. the method for claim 1, it is characterised in that：Impurity in tablet samples is determined with HPLC in step (2) Content, by external standard method with calculated by peak area.

3. the method for claim 1, it is characterised in that：Tablet samples are determined with slide measure first in step (3) Diameter and thickness, then determine its hardness with tablet hardness tester, then calculate the tensile strength of tablet according to formula (1)

.

4. the method for claim 1, it is characterised in that：In step (2) and (3), the rejecting of singular value can be by modeling Software automatic screening and rejecting, and according to spectroscopic data, reference value or prediction effect artificial screening and rejecting.

5. the method for claim 1, it is characterised in that：The parameter of collection unknown sample NIR-DRS in step (4) Value with consistent in step (1)；The preprocessing procedures of unknown sample respectively with consistent in step (2) and (3).