Robust freeze-drying process re-design of a legacy product based on risk analysis and design of experiments.

In this study, a QbD freeze-drying process re-design applied to a lyophilized injectable drug product is presented. The main objective was to assess the freeze-drying process robustness using risk analysis and a proper experimental design. First, the product's thermal fingerprint was characterized by thermal analysis and freeze-drying microscopy. Then, according to the output of the risk analysis, primary drying temperature and pressure were studied by a Doehlert DoE design with four responses; primary drying time, appearance, residual moisture content, and reconstitution time. Statistically significant MLR models were obtained for residual moisture content and primary drying time. In the latter, the temperature factor was the predominant factor to predict the duration of the primary drying stage. Two additional lab-scale batches were run to confirm the mathematical model predictions. Finally, optimal primary drying conditions (30 °C, 0.400 mbar) were selected to minimize the duration of the primary drying stage, while preserving the quality of the product. It was possible to set high temperature and pressure values because no collapse temperature was found during the thermal characterization of the product. Secondary drying temperature and time were defined based on the residual moisture content results. It was shown that secondary drying is robust between 30 °C and 50 °C and from 3 to 16 h. In conclusion, we were able to define a robust freeze-drying process which was further validated at an industrial scale with satisfactory results and approved by the health authorities in different countries around Europe.

Finding a reliable limit of detection in the NIR determination of residual moisture in a freeze-dried drug product.

The validation of an analytical method in the pharmaceutical industry follows strictly regulated guidelines. The introduction of multivariable calibration methods requires a revision of these recommendations, since some of them are contradictory regarding the limit of detection (LOD). This work compares the LOD values obtained using pseudounivariate and multivariate procedures in the PLS-NIR determination of residual moisture content (RMC) in a freeze-dried drug. As NIR has proved to be more precise than Karl-Fischer at low RMC values, LOD has been estimated by ordinary and by orthogonal least squares regression. The precision of the RMC determination in approx. 2000 industrial vials was used as an indirect evidence of the reliability of the LOD values obtained. The effect of reducing the number of calibration samples and increasing the RMC values have also been studied. No significant differences were observed using a number of calibration samples ≥ 20. Based on our findings, when the size of the calibration sample set is high and the range of RMC values is close to the limit, the LOD estimated with the ICH formula and using orthogonal regression should be recommended. If water content moves away, the ICH formula should be replaced by the LODOS equation as a practical, reliable and simple procedure.

A comprehensive study of a new versatile microchip device based liquid phase microextraction for stopped-flow and double-flow conditions.

A new geometry for a versatile microfluidic-chip device based liquid phase microextraction was developed in order to enhance the preconcentration in microfluidic chips and also to enable double-flow and stopped-flow working modes. The microchip device was combined with a HPLC procedure for the simultaneous determination of two different families as model analytes, which were parabens and non-steroidal anti-inflammatories (NSAIDs): Ethyl 4-hydroxybenzoate (Et-P), Propyl 4-hydroxybenzoate (Pr-P), Butyl 4-hydroxybenzoate (Bu-P), IsoButyl 4-hydroxybenzoate (iBu-P), salycilic acid (SAC), ketoprofen (KET), naproxen (NAX), diclofenac (DIC) and ibuprofen (IBU) in urine samples. The new miniaturized microchip proposed in this work allows not only the possibility of working in double-flow conditions, but also under stagnant conditions (stopped-flow) (SF-µLPME). The sample (pH 1.5) was delivered to the SF-µLPME at 20 µL min-1 while keeping the acceptor phase (pH 11.75) under stagnant conditions during 20 min. The highest enrichment factors (between 16 and 47) were obtained under stopped-flow conditions at 20 µL min-1 (sample flow rate) after 20 min extraction; whereas the extraction efficiencies were within the range of 27-81% for all compounds. The procedure provided very low detection limits between 0.7 and 8.5 µg L-1 with a sample volume consumption of 400 µL. Parabens and NSAIDs have successfully been extracted from urine samples with excellent clean up and recoveries over 90% for all compounds. In parallel, the new device was also tested under double flow conditions, obtaining good but lower enrichment factors (between 9 and 20) and higher extraction efficiencies (between 45 and 95) after 7 min extraction, consuming a volume sample of 140 µL. The versatile device offered very high extraction efficiencies and good enrichment factor for double flow and stopped-flow conditions, respectively. In addition, this new miniaturized SF-µLPME device significantly reduced costs compared to the existing analytical techniques for sample preparation since this microchip require few microliters of sample and reagents and it is reusable.

Raman spectroscopy for the analytical quality control of low-dose break-scored tablets.

Quality control of solid dosage forms involves the analysis of end products according to well-defined criteria, including the assessment of the uniformity of dosage units (UDU). However, in the case of break-scored tablets, given that tablet splitting is widespread as a means to adjust doses, the uniform distribution of the active pharmaceutical ingredient (API) in all the possible fractions of the tablet must be assessed. A general procedure to accomplish with both issues, using Raman spectroscopy, is presented. It is based on the acquisition of a collection of spectra in different regions of the tablet, that later can be selected to determine the amount of API in the potential fractions that can result after splitting. The procedure has been applied to two commercial products, Sintrom 1 and Sintrom 4, with API (acenocoumarol) mass proportion of 2% and 0.7% respectively. Partial Least Squares (PLS) calibration models were constructed for the quantification of acenocoumarol in whole tablets using HPLC as a reference analytical method. Once validated, the calibration models were used to determine the API content in the different potential fragments of the scored Sintrom 4 tablets. Fragment mass measurements were also performed to estimate the range of masses of the halves and quarters that could result after tablet splitting. The results show that Raman spectroscopy can be an alternative analytical procedure to assess the uniformity of content, both in whole tablets as in its potential fragments, and that Sintrom 4 tablets can be perfectly split in halves, but some cautions have to be taken when considering the fragmentation in quarters. A practical alternative to the use of UDU test for the assessment of tablet fragments is proposed.

Fast assessment of the surface distribution of API and excipients in tablets using NIR-hyperspectral imaging.

The inclusion of hyperspectral imaging systems in the manufacturing and development of pharmaceutical products is allowing a successful improvement in the quality control of solid dosage forms. The correct distribution not only of active pharmaceutical ingredient (API) but also of the rest of excipients is essential to assure the correct behavior of the tablet when ingested. This is especially relevant in tablets with low content of potent APIs, in which the prescribed intake dosage frequently corresponds to half-a-tablet. Therefore, the aim of this work is to study the surface distribution of the compounds in tablets with low API content. The proposed procedure includes the scanning of the tablet surface using near infrared hyperspectral spectroscopy in association with multivariate curve resolution (MCR) techniques to obtain selective pictures for each individual compound and to allow the fast assessment of their distribution in the measured surface. As an example, a set of commercial Lorazepam tablets (approximately 1% mass fraction of API, and four excipients) were analyzed. The results obtained show the capacity of the proposed methodology as an expedite approach to evaluate the uniformity of the contents between and within tablets. A method to estimate the homogeneity distribution of API in the two halves of the tablet is also proposed.

Implementation of enhanced correlation maps in near infrared chemical images: application in pharmaceutical research.

Recent developments in Hyperspectral Imaging equipment have made possible the use of this analytical technique for fast scanning of sample surfaces. This technique has turned out to be especially useful in Pharmacy, where information about the distribution of the components in the surface of a tablet can be obtained. One particular application of Hyperspectral Chemical Imaging is the search for singularities inside pharmaceutical tablets, e.g. coating defects. Nevertheless, one problem has to be faced: how to analyze a sample without any previous knowledge about it, or having only the minimum information about the tablet. In this work a new methodology, based on correlation coefficients, is introduced to obtain valuable information about one Hyperspectral Image (detection of defects, punctual contaminants, etc.) without any previous knowledge. The methodology combines Principal Component Analysis (PCA), correlation coefficient between one specific pixel included in the image and the rest of the image; and a new enhanced contrast function to obtain more selective chemical and spatial information about the image. To illustrate the applicability of the proposed methodology, real tablets of ibuprofen have been studied. The proposed methodology is presented as a control technique to detect batch variability, defects in final tablets and punctual contaminants, being a potential supplementary tool for quality controls. In addition, the usefulness of the proposed methodology is not exclusive to NIR-CI devices, but to any hyperspectral and multivariate image system.

Application of representative layer theory to near-infrared reflectance spectra of powdered samples.

The diffuse reflectance near-infrared (NIR) spectrum of a powdered sample includes the contribution of specular and diffuse reflectance, which is a function of absorbance and scattering. The fraction of light scattered depends in a complex manner on the physical properties of the sample such as particle size, refraction index, etc. Several theories to study the dependence of NIR spectra on the particle size have been proposed. The best known is the Kubelka-Munk model, an approach based on continuous mathematics. Recently Dahm and Dahm put forward an alternative method, the representative layer theory (RLT), which uses discontinuous mathematics as a basis. This approach can be used to identify and disentangle the scattering and absorbance signals as well as their dependence on the particle size. The scattering and absorption coefficient of NaCl (a nonabsorbing material) and of potassium hydrogen phthalate, KHP (a strong absorber), have been estimated through the application of the representative layer theory, working on a particle size range from 63 to 450 microm. In both samples, the absorption coefficient of the sample (K) remains constant and practically independent of the particle size, while the scattering coefficient of the sample (S) decreases when the particle diameter increases, becoming stable around a diameter of 250 microm.

Optimization of the xylan degradation activity of monolithic enzymatic membranes as a function of their composition using design of experiments.

The aim of this work was the development and optimization of enzymatic monolithic membranes with high catalytic activity for the degradation of xylan into xylooligosaccharides. The chemometric tool design of experiments has been utilized here for the first time for the optimization of the enzymatic activity of the monolithic membranes based on their constituents. The effect of three process variables, including the amount of various monomer contents and the porogenic solvents ratio, has been studied on the enzymatic activity of the resulted membranes. The experimental design chosen was a central face centred with six central points in order to obtain an orthogonal model, with the precision of the results being independent of the range of values considered for each parameter. The software Modde(c) 6.0 from Umetrics(c) was used to build and analyze the results of the experimental design using partial least squares regression. The optimization of the suggested model provided the best membrane composition to achieve maximum enzymatic activity, which can be related to the amount of enzyme immobilized on the monolithic membrane. The predictive capacity of the model was evaluated performing additional experiments.

Parallel factor analysis combined with PLS regression applied to the on-line monitoring of Pichia pastoris cultures.

Various key variables (biomass, substrate and product) of bioprocesses should be monitored in order to retrieve useful information on the system, with the biomass (the cell density) the principal target. Although several analytical methods have been adapted and used to monitor the evolution of cell density evolution in cultures, a general method for performing this determination has not yet been established, as each technique has its own advantages and drawbacks. In the present work, noninduced glycerol batch cultures (for which biomass and substrate are the key variables) were monitored using multiwavelength fluorescence spectroscopy. The data gathered were modelled via PARAFAC-PLS chemometric methodologies, resulting in important qualitative and quantitative information about the behaviours of different biogenic fluorophors in batch cultures of the yeast Pichia pastoris. This information was used to predict the target process variables in such cultures; this permitted the applicability of this combined technique to bioprocess monitoring to be assessed.

Three-way partial least-squares regression for the simultaneous kinetic-enzymatic determination of xanthine and hypoxanthine in human urine.

The performance of three-way principal component analysis and three-way partial least-squares regression when applied to a complex kinetic-enzymatic system is studied, in order to investigate the analytical potential of the combined use of these chemometric technologies for non-selective enzymatic systems. A enzymatic-kinetic procedure for the simultaneous determination of hypoxanthine and xanthine in spiked samples of human urine is proposed. The chemical system involves two consecutive reactions catalyzed by xanthine oxidase (EC 1.17.3.2). This enzyme catalyzes the oxidation of hypoxanthine, first to xanthine and then to uric acid, a competitive inhibitor of the reactions. The influence of uric acid during quantitative determination was considered in the design of the calibration set. The sample and enzyme solution were mixed in a stopped-flow module and the reaction was monitored using a diode array spectrophotometer. The recorded data have an intrinsical three-component structure (samples, time and wavelength). This data array was studied via three-way principal component analysis and was modeled for quantitative purposes using a three-way partial least-squares calibration procedure. Results are compared with those obtained by applying classical bilinear PLS to the previously unfolded data matrix.