No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
CHEMOMETRICS IN PHARMACEUTICAL ANALYSIS
Abstract
In this introduction to the subject of chemometrics, attention is focused on experimental design, considering both manual and automated applications of chemometric methods to systems of pharmaceutical interest. Additionally, the use of robotics and expert systems is briefly discussed.
Background:
To ensure quality and stability, monitoring systems are recommended to analyze pharmaceutical manufacturing processes.
Objective:
This study was performed to predict powder X-ray diffraction (PXRD) patterns of formulation powders through attenuated total reflectance (ATR)-infrared (IR) spectroscopy in a nondestructive manner along with chemometrics.
Results:
Caffeine anhydrate, acetaminophen, and lactose monohydrate were grinded at six weight ratios. The six sample groups were evaluated using ATR-IR spectroscopy and PXRD analysis. Partial least squares models were constructed to predict the PXRD intensities of the samples from the ATR-IR spectra. The prediction accuracy on the prepared PLS regression models was as high as R2 = 0.993.
Conclusions:
Linear relationships were obtained between the prediction data set and reference PXRD intensity at each degree. 2D PLS regression coefficient analysis enabled the analysis of the correlation between PXRD patterns and IR spectra.
Salt and cocrystal formulations are widely used as techniques to improve physicochemical properties of compounds. Some spectrometric techniques to distinguish cocrystals from salts have been reported; however, it has not been possible to adapt these formulations with many compounds, because of limitations, high difficulty, and exceptions. Therefore, we focused on the possibility of UV spectrometry, which had not been reported as a classification technique for salts and cocrystals. The integration values of solid-state UV/visible (Vis) spectra of indomethacin salts were larger than those of physical mixtures of indomethacin and counter molecules, while that of indomethacin cocrystal was not large compared with that of the physical mixture. From these results, differences between a salt and a cocrystal were observed in their solid-state UV/Vis absorption spectra for indomethacin complexes. Therefore, it is suggested that solid-state UV/Vis absorption spectra can be used as a new technique to classify salts and cocrystals.
Graphical Abstract Fullsize Image
Application of design of experiments to study the robustness of a HPLC method was carried out, where six factors were selected: the percentage of acetonitrile (ACN) in the mobile phase, the pH of mobile phase, the detector wavelength, the column temperature, the flow rate and the strength of the buffer. These factors are examined in two-level screening experimental designs created using JMP@ (SAS Institute) software. The system suitability parameters like the capacity factor, tailing factor, resolution were calculated and analyzed using the ANOVA method by least squares fitting. The results are within the acceptance criteria showing that the method is robust within the established limits. Forced degradation studies were also conducted for sitagliptin and pioglitazone.
Chemometrics is now a well established discipline prompting independent research and development in its own right but, for their continuing success, chemometric developments must be relevant to the real needs of analysis. Pharmaceutical analysis, i.e., physical and chemical assessment of drugs and their dosage forms, needs more formal application of mathematical and statistical methods. The organisation of the laboratory, the development, optimization, assessment and interpretation of methods, and the evaluation of the data produced can all benefit from the application of chemometrics. Here, a selection of the more challenging problems facing the pharmaceutical analyst is presented, the potential for chemometrics is considered and some consequent implications for utilisation, against a background of control and regulation are discussed.
Chemometrics is playing an increasingly important role in flavor research. Pattern recognition techniques based on multivariate analysis have been most useful in processing data from chromatography and spectrometry mainly due to the intrinsic multi dimensionality of flavor. Multiple regression analysis and its derivatives including partial least squares regression (PLS) have been frequently used for correlating instrumental data to sensory properties. Factor analysis and principal component analysis are widely used for searching latent factors and extracting information as unsupervised pattern recognition. Cluster analysis and discriminant analysis have been successful for classification of samples; however, modeling of samples using SIMCA and nonparametric classification such as KNN have also gained popularity for improving accuracy. Simplex optimization has been well established as a technique in chemometrics, however, it is relatively unknown in flavor research. Computer‐aided optimization has a great potential for application to flavor study, for example, operating conditions, reconstruction, and blending. Fuzzy theory, artificial intelligence, and robotics will become important methodologies in flavor research in the future.
The chemometrics approach has been applied to the analysis of the combined effects of temperature, light intensity and CO2 concentration on the growth rate of the blue-green alga Anacystis nidulans. The strategy for the experiments is based on an extended factorial design. A statistically significant interaction is found between temperature and light intensity. Optimal conditions for maximal growth rate can be predicted by multiple regression analysis and its reliability has been confirmed by experiments. At the optimal growth condition the photosynthetic nature of A. nidulans was examined. The results suggest the importance of the utilization of these cells in the logarithmic growth phase for the purpose of CO2 fixation.
Abstract The possibilities of employing methods of chemometrics in order to characterize macromolecules are described. The review has been limited to chemometric methods concerning multivariate data analysis. Principal component analysis (PCA) has shown to be very useful for pattern recognition problems arising from chromatographic and spectroscopic data. An example of using a classification technique, SIMCA (Soft Independent Modelling of Class Analogy), as a product control method is presented. The suitability of Partial Least Squares (PLS) for relating data of different natures, e.g. chemical data with biological data, is discussed. Moreover, examples ranging from spectroscopic determinations to QSAR (Quantitative Structure Activity Relationships) are illustrated.
IntroductionResponse Surfaces, Sampling, and OptimizationSignal ProcessingPrincipal Components Analysis and Factor AnalysisCalibration and Mixture AnalysisClassification and ClusteringLibrary SearchingMolecular Structure-Property RelationshipsGas Chromatographic Retention Indices for Diverse Drug CompoundsSimulation of Carbon-13 Nuclear Magnetic Resonance Spectra of Methyl-Substituted Norbornan-2-olsSummary and Conclusions
Plackett-Burman experimental design has been studied in the chiral separation of clenbuterol using capillary electrophoresis (CE). This saturated fractional design approach was used to simultaneously investigate the variables pH, cyclodextrin concentration, electrolyte ionic strength, methanol concentration, and injection time, each at three levels. Percentage main effects of each variable on resolution, retention time and peak efficiency were calculated. Ionic strength, pH and -cyclodextrin concentration were found to exert the greatest effect on resolution, pH on retention time and pH and ionic strength on peak efficiency. Results are in agreement with those from a previous univariate study. It is concluded that Plackett-Burman experimental design offers of rapid means for testing robustness of chiral CE methods or for the simultaneous screening of experimental variables.
Qualitative and quantitative analysis of the essential oil of Thymus willkomii showed linalool, α-terpinyl acetate and 1,8-cineole as major constituents. Chemometric investigation of the infraspecific variability of the essential oil of a population of this species led to the differentiation of two main chemotypes; type A: linalool (49% of the plants) and type B: α-terpinyl acetate (36% of the plants). The remaining plants (15%) could not be included in either of these two types; some of them showed significant amounts of other constituents such as trans-sabinene hydrate, myrcen-8-ol, and terpinen-4-ol.
The purpose of this research is to apply near infrared spectrometry (NIR) with chemoinformetrics to predict the change of crystalline properties of indomethacin (IMC) amorphous under various levels of relative humidity storage conditions. Stability test for amorphous and meta-stable polymorphic forms was performed in humidity controlled the modified 96-well quartz plates containing various kinds of saturated salt solutions (0-100% of relative humidity (RH)) by NIR spectroscopy. Amorphous form was obtained melt product to pour into liquid nitrogen and after then ground. Samples were stored at 25°C in the 6-well plates at various levels of RH. The spectra of the powder samples were measured by the reflectance FT-NIR spectrometer. The second derivative spectra of form α showed specific absorption peaks at 4980, 6036, 7296 and 8616 cm-1 and that of form γ showed those at 5020, 5028, 7344, 7428 and 8436 cm-1. After storage at less than 50% RH, the peak intensities at 5020, 5028, 7344, 7428 and 8436 cm-1 of the amorphous solid increased with increasing of storage time. However, the peak intensity at 4980, 6036 and 7296 cm-1 increased at more than 50% RH Please check and confirm the edit. The results suggested that at lower humidity, the IMC amorphous solid transformed into form γ, but it transformed into form α at more than high humidity. It is possible that crystalline stability of the pharmaceutical preparations could be predicted by using humidity controlled 96-well plates and reflectance NIR-chemoinformetric methods.
The purpose of the present study is to measure polymorphic content in a bulk powder, mefenamic acid polymorph of pharmaceuticals, as a model drug by THz-spectrometer using frequency-tunable THz wave generators based on difference-frequency generation in gallium phosphate crystals. Mefenamic acid polymorphic forms I and II were obtained by recrystallisation. Eleven standard samples varying a various polymorphic form I content (0-100%) were prepared by physical mixing. After smoothing and area normalising, the THz-spectra of all standard samples showed an isosbestic point at 3.70 THz. After the THz-spectral data sets were arranged into five frequency ranges, and pretreated using various functions, calibration models were calculated by the partial least square regression method. The effect of spectral data management on the chemometric parameters of the calibration models was investigated. The relationship between predicted and actual form I content was the best linear plot. On the regression vector (RV) that corresponded to absorption THz-spectral data, the peak at 1.45 THz was the highest value, and the peak at 2.25 THz was the lowest on RV. THz-spectroscopy with chemometrics would be useful for the quantitative evaluation of mefenamic acid polymorphs in the pharmaceutical industry. This method is expected to provide a rapid and nondestructive quantitative analysis of polymorphs.
Since polymorphs exhibit differences in chemical and physicochemical stability, characteristics, and dissolution rate of the bulk powder, they may significantly affect on the bioavailability of pharmaceutical compounds.
The purpose of the present study is to establish a method for determining the carbamazepine (CBZ) polymorphic content of a double-layered tablet containing various ratios of forms I and III by using transmittance- and reflectance-near-infrared (TNIR and RNIR) spectroscopy involving chemometrics.
Both TNIR and RNIR instruments were used to analyze both top (form I) and wire (form III) sides of the compacts, respectively. NIR spectra were analyzed to predict polymorphic content by a principal component regression analysis. NIR data of the tablets were divided into two wavelength ranges: between 860 and 1680 nm (FW), and 1245 and 1285 nm (NW).
The calibration models for polymorphic content based on TNIR had a linear relationship, but those based on RNIR did not. The accuracy of the calibration models suggested that the double-sided data set is more robust than the single-sided data set. Since the spectra of FW involved various information, the calibration models showed a linear correlation, but it is difficult to understand their model. In contrast, those of NW provided limited information on polymorphic forms making it very easy to understand the model.
Limiting the wavelength of the spectra is useful to help understand the calibration-complicated model.
The development of a new drug substance and its dosage forms demands the establishment and implementation of suitable analytical methods. Through the application of chemometrics, "intelligent" laboratory systems and integrated data handling systems, considerable progress is being made in automating both the development and routine use of analytical methods in pharmaceutical analysis.
The ideal laboratory robot can be viewed as "an indefatigable assistant capable of working continuously for 24 h a day with constant efficiency". The development of a system approaching that promise requires considerable skill and time commitment, a thorough understanding of the capabilities and limitations of the robot and its specialized modules and an intimate knowledge of the functions to be automated. The robot need not emulate every manual step. Effective substitutes for difficult steps must be devised. The future of laboratory robots depends not only on technological advances in other fields, but also on the skill and creativity of chromatographers and other scientists. The robot has been applied to automate numerous biomedical chromatography and electrophoresis methods. The quality of its data can approach, and in some cases exceed, that of manual methods. Maintaining high data quality during continuous operation requires frequent maintenance and validation. Well designed robotic systems can yield substantial increase in the laboratory productivity without a corresponding increase in manpower. They can free skilled personnel from mundane tasks and can enhance the safety of the laboratory environment. The integration of robotics, chromatography systems and laboratory information management systems permits full automation and affords opportunities for unattended method development and for future incorporation of artificial intelligence techniques and the evolution of expert systems. Finally, humanoid attributes aside, robotic utilization in the laboratory should not be an end in itself. The robot is a useful tool that should be utilized only when it is prudent and cost-effective to do so.
The purpose of this research was to develop a rapid chemometrical method based on near-infrared (NIR) spectroscopy to determine indomethacin (IMC) polymorphic content in mixed pharmaceutical powder and tablets. Mixed powder samples with known polymorphic contents of forms alpha and gamma were obtained from physical mixing of 50% of IMC standard polymorphic sample and 50% of excipient mixed powder sample consisting of lactose, corn starch, and hydroxypropylcellulose. The tablets were obtained by compressing the mixed powder at 245 MPa. X-ray powder diffraction profiles and NIR spectra were recorded for 6 kinds of standard materials with various polymorphic contents. The principal component regression analysis was performed based on normalized NIR spectra sets of mixed powder standard samples and tablets. The relationships between the actual and predicted polymorphic contents of form g in the mixed powder measured using x-ray powder diffraction and NIR spectroscopy show a straight line with a slope of 0.960 and 0.995, and correlation coefficient constants of 0.970 and 0.993, respectively. The predicted content values of unknown samples by x-ray powder diffraction and NIR spectroscopy were reproducible and in close agreement, but those by NIR spectroscopy had smaller SDs than those by x-ray powder diffraction. The results suggest that NIR spectroscopy provides a more accurate quantitative analysis of polymorphic content in pharmaceutical mixed powder and tablets than does conventional x-ray powder diffractometry.
A new analytical system that uses the rotational signatures of gas phase molecules is described and demonstrated. It uses a solid state source to probe molecular systems in the millimetre and submillimetre wave range, the only region of the electromagnetic spectrum not yet used extensively for analytical purposes. It employs the FAst Scan Submillimetre Spectroscopy Technique (FASSST), which leads to an especially simple system architecture. Among the attributes of the system are generality, sensitivity, 'absolute' specificity, small size, simplicity, and the potential for very low cost. Applications to problems of analytical interest are also discussed.
A Diltiazem kinetic spectrophotometric method was optimized by factorial analysis. The experimental method is based on a two-stage reaction of Diltiazem with hydroxylamine and a ferric salt: in the first stage there is a hydroxamic acid formation; and, in the second stage there is a red colour complex ferric hydroxamate formation. The variables under investigation were: solvent; hydroxylamine, sodium hydroxide and ammonium ferric sulphate concentrations; volume of perchloric acid; and, temperature. The responses of the reactional system were the maximum absorbance, the wavelength and the reaction time at maximum absorbance. Experimental design methodologies were used in the optimization. Fractional and full factorial designs followed by optimization Box-Behnken and central composite experimental designs were used. The observed optimum conditions were: methanol as reaction solvent; hydroxylamine concentration of 9.375%; sodium hydroxide concentration of 18.750%; ferric reagent concentration of 2.000%; minimum volume of perchloric acid to neutralize the sodium hydroxide; and, room temperature as reaction temperature. With this set of experimental conditions a reaction time of 10.5s with maximum colour development at 512nm wavelength was achieved.
The reversed-phase chromatography of sulindac and its structurally related compounds was investigated. Several mobile phase parameters were examined. The best resolution was obtained using a mixed organic modifier consisting of methanol and acetonitrile. Employing a factorial experimental design, the effect of four primary operating variables was determined and a computer simulation of the chromatographic process was carried out. This adequately accounted for the interactive effects, accurately predicted optimum conditions and indicated corrective changes in the operating parameters to obtain a desired separation.With the simulated model, further information concerning column performance was available. The enthalpy of binding of each component at various mobile phase conditions was calculated and the effect of temperature on column efficiency documented. The predicted optimum in column efficiency at 35–45°C could be explained, in part, by the kinetic and diffusional processes occurring within the microparticulate high-performance liquid chromatographic column.
Atomic absorption (AA) methods of analysis are subject to the influence of many experimental factors such as fuel and oxidant flow rates, burner height, and lamp current. The optimization of these factors and the corresponding improvement of response are integral parts of the development of AA methods. The importance of the systematic improvement of response was recognized by Cellier and Stace, who used sequential factorial designs to increase analytical sensitivity. Parsons and Winefordner considered the theoretical significance of various experimental factors and also employed factorial designs to achieve improved response. Other researchers have used similar statistical approaches in the development of AA methods.
Chemometrics is the chemical discipline that uses mathematical and statistical methods to design or select optimal measurement procedures and experiments and to provide maximum chemical information by analyzing chemical data. This article describes the impact of chemometrics on analytical chemistry with the example of chromatography. It focuses on applications of information theory, experimental optimization and pattern recognition.
The usefulness of two standardized HPLC-systems for the analysis of basic drugs in tablets and capsules is exemplified. The standardized HPLC systems both use a CN-column combined with either a polar or a non polar mobile phase. Also the sample preparationis standardized and simple; it involves suspension of the powder mixture in one of the mobile phase components, centrifugation and injection of the clear supernatant.
The trend in analytical chemistry toward statistically designed experiments is reinforced by the trend toward using personal computers. The tedious calculations required to determine which factors significantly affect the performance or yield of an experiment are easily done with one of the 'electronic spreadsheet' programs now available on personal computers.
The success of chemometric techniques that operate on data obtained from experiments is often considerably improved if the data have been acquired by a systematic design. The intimate relationships among good chemometric results and good models, good measurement processes, and good experimental designs are shown. A bibliography of the applications of sequential simplex optimization in chemometrics is presented.
FT-n.m.r. spectra were automatically phase-corrected using the modified simplex method. Three optimization criteria for phase correction were investigated. Best results were obtained by maximization of the intensity minimum and maximization of the summed intensities below the baseline. The maximization of spectral area consistently failed to correct the spectral data satisfactorily.
It is shown that the modified sequential simplex algorithm can be used as the basis for the unattended optimisation of reversed-phase liquid chromatographic separations. A chromatographic response function is computed to evaluate individual chromatograms with respect to resolution and analysis time. This function is used automatically by a microprocessor-controlled chromatograph to optimise experimental parameters. The use of the procedure is illustrated by the automated optimisation of two and three component mobile phases and the optimisation of a gradient elution profile.
The separation of morphine, codeine, noscapine and papaverine in reversed-phase ion-pair chromatography has been investigated by means of statistical optimization methods. The value of the capacity factor, as a function of the methanol—water ratio, pH and the concentrations of buffer and ion-pair reagent, including their interactions, was studied for each compound. It was shown that, by using such methods, conditions could be chosen which essentially improved the separation and that a quantitative characterization of the optimum region was facilitated.
A methodology for the development of separation procedures for different groups of substances (mostly drugs) to described. Chromatography is performed on a cyanopropyl column with a limited number of mobile phase solvents to carry out normal-phase as well as reversed-phase HPLC. Initially, a gradient edition with two solvents is performed from which a starting isocratic mobile phase composition is selected to elute the substances in a suitable capacity factor range. Other binary and ternary solvent compositions, determined with an experimental design and having the same solvent strength but different selectivity, are tested to adjust the resolution of peak pairs and to improve the separation between the solutes. Benzamides, local anesthetics, phenothlazines, tricyclic antidepressants, benzodiazepines, corticosterolds, sulfonamides, barbiturates, and food preservatives are separated either in normal phase or in reversed phase or in both.
The combined effects of pH (3.6–6.0) and octylamine hydrochloride concenttration (0–5.0 mM) in methanol—water (20:80) eluents were determined for hydrocinnamic, trans-cinnamic, phenylacetic, trans-p-coumaric, trans-ferulic, trans-caffeic and vanillic acids; for phenylethylamine; and for phenylalanine. The effects are described by a simple ion-interaction model that does not require ion-pair formation in either phase and is not based upon classical ion exchange. The simplicity and generality of the mathematical forms of the model make it useful for predicting retention behavior.
The application of chemometrics to extract chemical information from analytical measurements will ensure that full advantage is taken of computerized instrumentation and that analytical chemistry evolves as a true information science.
The usefulness of amperometric detection in pharmaceutical analyses was investigated for different groups of drugs. The UV response at 254 nm and that at the absorption maximum of the solute were compared with the electrochemical signal obtained. The minimum detectable concentration (nanograms on-column) of each substance is reported for the three different detection systems. This comparison was performed for 72 drugs (local anaesthetics, antipyretics, tricyclic antidepressants, sulphonamides, sex hormones, beta-adrenoceptor blocking agents, phenothiazines, alkaloids, diuretics and penicillins). The median limit of detection of the amperometric detector (see definition in the text) is 1.0 ng on-column and the median gain in sensitivity, compared with UV detection is 22.5.
Assay methods for measuring saliva levels of carbamazepine and its active metabolite, and of amidopyrine are developed using a standardized analysis strategy. Both assay methods include ion-pair extraction of the analytes with octylsulphate as the counter-ion and chromatography on a CN-bonded phase using a hexane--dichloromethane--acetonitrile--propylamine mixture as the mobile phase. Both methods are applied to patient samples.
The composition of pharmaceutical formulations is often subject to trial and error. This approach is time consuming and unreliable in finding the best formulation. Optimization by means of an experimental design might be helpful in shortening experimenting time. Such a design with the concomitant mathematical models, reveals effects and interactions of the variables. The independent variables are the different compositions of the mixtures of the chosen ingredients [drug(s) and excipients]. The dependent variables are the properties (responses) of the formulation. When all responses of interest have been expressed in models that describe the response as a function of the composition of the mixture, the models can be combined graphically or mathematically to find a composition satisfying all demands. In this paper an introduction to the use of mixture designs will be given by means of a theoretical part and an example: optimizing a tablet formulation consisting of excipients only.
An approach to analyzing and interpreting kinetic data from stability studies using factorial designs is presented. This may be useful for screening purposes or as an aid in identifying significant effects in complex systems. A typical 2n factorial experiment is discussed, and methods of variance estimation and statistical testing are presented. An example of simulated data is used to demonstrate how typical results may be analyzed, as well as the potential and limitations of this design in interpretation and construction of kinetic models.
A test set of 100 basic drugs has been chromatographed on 16 preselected HPLC-systems using four different types of stationary phase (Si-, NH2-, CN- and C18-). A numerical treatment of the chromatographic data, based on the discriminating power concept, results in the selection of two preferred HPLC systems for basic drugs. both using the CN-bonded phase. The preferred eluents are n-heptane-dichloromethane-acetonitrile-propylamine (25:50:25:0.1) and acetonitrile-water-propylamine (90:10: 0.01). The two preferred HPLC systems are adopted in a standardized analysis strategy.
The first of a series of papers on the development of a standardized analysis strategy for basic drugs explains the possible advantages and philosophy of the strategy. The scheme uses ion-pair extraction with direct injection of the extracts into an HPLC system emanating, from two previously-selected systems. The extraction efficiency of sodium-n-octylsulphate as the ion-pairing reagent is compared with that of di(2-ethylhexyl) phosphoric acid (previously shown to be generally applicable to the extraction of basic drugs). Direct injection of the ion-pair extracts into an HPLC system is possible because retention behaviour is independent of whether the basic drugs are injected as an ion-pair or as a base.
The feasibility of using expert systems for the development of analytical procedures is investigated. A system for the computer generation of procedures to determine active drug substances in commercial formulations is proposed. It is shown that in nearly 85% of the cases investigated the present system immediately yields a correct procedure or conclusion. It is concluded that selecting methods and developing procedures with the use of expert systems is difficult but feasible.