The adsorption of the proton-pump inhibitor omeprazole was investigated using RP-LC with chemometric models combined with adsorption isotherm modelling to study the effect of pH and type of organic modifier (i.e., acetonitrile or methanol). The chemometric approach revealed that omeprazole was tailing with methanol and fronting with acetonitrile along with increased fronting at higher pH. The increased fronting with higher pH for acetonitrile was explored using a pH-dependent adsorption isotherm model that was determined using the inverse method and it agreed well with the experimental data. The model indicated that the peaks exhibit more fronting at high pH due to a larger fraction of charged omeprazole molecules. This model could accurately predict the shape of elution profiles at arbitrary pH levels in the studied interval. Using a two-layer adsorption isotherm model, the difference between acetonitrile and methanol was studied at the lowest pH at which almost all omeprazole molecules are neutral. Omeprazole had adsorbate–adsorbate interactions that were similar in strength for the acetonitrile and methanol mobile phases, while the solute–adsorbent interactions were almost twice as strong with methanol. The difference in the relative strengths of these two interactions likely explains the different peak asymmetries (i.e., tailing/fronting) in methanol and acetonitrile. In conclusion, thermodynamic modelling can complement chemometric modeling in HPLC method development and increase the understanding of the separation.

This is the first investigation in a series that aims to enhance the scientific knowledge needed for reliable analytical method transfer between HPLC and UHPLC using the quality by design (QbD) framework. Here, we investigated the differences and similarities from a thermodynamic point of view between RP-LC separations conducted with 3.5 μm (HPLC) and 1.7 μm (UHPLC) C18 particles. Three different model solutes and one pharmaceutical compound were used: the uncharged cycloheptanone, the cationic benzyltriethylammonium chloride, the anionic sodium 2-naphatlene sulfonate and the pharmaceutical compound omeprazole, which was anionic at the studied pH. Adsorption data were determined for the four solutes at varying fractions of organic modifier and in gradient elution in both the HPLC and UHPLC system, respectively. From the adsorption data, the adsorption energy distribution of each compound was calculated and the adsorption isotherm model was estimated. We found that the adsorption energy distribution was similar, with only minor differences in degree of homogeneity, for HPLC and UHPLC stationary phases. The adsorption isotherm model did not change between HPLC and UHPLC, but the parameter values changed considerably especially for the ionic compounds. The dependence of the organic modifier followed the same trend in HPLC as in UHPLC. These results indicates that the adsorption mechanism of a solute is the same on HPLC and UHPLC stationary phases which simplifies design of a single analytical method applicable to both HPLC and UHPLC conditions within the QbD framework.

The importance of the generated temperature and pressure gradients in ultra-high-pressure liquid chromatography (UHPLC) are investigated and compared to high-pressure liquid chromatography (HPLC). The drug Omeprazole, together with three other model compounds (with different chemical characteristics, namely uncharged, positively and negatively charged) were used. Calculations of the complete temperature profile in the column at UHPLC conditions showed, in our experiments, a temperature difference between the inlet and outlet of 16 degrees C and a difference of 2 degrees C between the column center and the wall. Through van't Hoff plots, this information was used to single out the decrease in retention factor (k) solely due to the temperature gradient. The uncharged solute was least affected by temperature with a decrease in k of about 5% while for charged solutes the effect was more pronounced, with k decreases up to 14%. A pressure increase of 500 bar gave roughly 5% increase in k for the uncharged solute, while omeprazole and the other two charged solutes gave about 25, 20 and 15% increases in k, respectively. The stochastic model of chromatography was applied to estimate the dependence of the average number of adsorption/desorption events (n) and the average time spent by a molecule in the stationary phase (tau(s)) on temperature and pressure on peak shape for the tailing, basic solute. Increasing the temperature yielded an increase in n and decrease in tau(s) which resulted in less skew at high temperatures. With increasing pressure, the stochastic modeling gave interesting results for the basic solute showing that the skew of the peak increased with pressure. The conclusion is that pressure effects are more pronounced for both retention and peak shape than the temperature effects for the polar or charged compounds in our study. (C) 2015 Elsevier B.V. All rights reserved.

Quality Control methods (QC-methods) play an important role in the overall control strategy for drug manufacturing. However, efficient life-cycle management and continual improvement are hindered due to a variety of post-approval variation legislations across territories and a lack of harmonization of the requirements. As a result, many QC-methods fall behind the technical development. Developing the QC-method in accordance with the Quality by Design guidelines gives the possibility to do continual improvements inside the original Method Operable Design Region (MODR). However, often it is necessary to do changes outside the MODR, e.g. to incorporate new technology that was not available at the time the original method was development. Here, we present a method enhancement concept which allows minor adjustments, within the same measuring principle, outside the original MODR without interaction with regulatory agencies. The feasibility of the concept is illustrated by a case study of a QC-method based on HPLC, assumed to be developed before the introduction of UHPLC, where the switch from HPLC to UHPLC is necessary as a continual improvement strategy. The concept relies on the assumption that the System Suitability Test (SST) and failure modes are relevant for other conditions outside the MODR as well when the same measuring principle is used. It follows that it should be possible to move outside the MODR as long as the SST has passed. All minor modifications of the original, approved QC-method must be re-validated according to a template given in the original submission and a statistical equivalence should be shown between the original and modified QC-methods. To summarize, revalidation is handled within the pharmaceutical quality control system according to internal change control procedures, but without interaction with regulating agencies.

Large pressure gradients are generated in ultra-high-pressure liquid chromatography (UHPLC) using sub–2 μm particles causing significant temperature gradients over the column due to viscous heating. These pressure and temperature gradients affect retention and ultimately result in important selectivity shifts. In this study, we developed an approach for predicting the retention time shifts due to these gradients. The approach is presented as a step-by-step procedure and it is based on empirical linear relationships describing how retention varies as a function of temperature and pressure and how the average column temperature increases with the flow rate. It requires only four experiments on standard equipment, is based on straightforward calculations, and is therefore easy to use in method development. The approach was rigorously validated against experimental data obtained with a quality control method for the active pharmaceutical ingredient omeprazole. The accuracy of retention time predictions was very good with relative errors always less than 1% and in many cases around 0.5% (n = 32). Selectivity shifts observed between omeprazole and the related impurities when changing the flow rate could also be accurately predicted resulting in good estimates of the resolution between critical peak pairs. The approximations which the presented approach are based on were all justified. The retention factor as a function of pressure and temperature was studied in an experimental design while the temperature distribution in the column was obtained by solving the fundamental heat and mass balance equations for the different experimental conditions. We strongly believe that this approach is sufficiently accurate and experimentally feasible for this separation to be a valuable tool when developing a UHPLC method. After further validation with other separation systems, it could become a useful approach in UHPLC method development, especially in the pharmaceutical industry where demands are high for robustness and regulatory oversight.

A fundamental investigation of the pressure effect on individual adsorption sites was undertaken based on adsorption energy distribution and adsorption isotherm measurements. For this purpose, we measured adsorption equilibrium data at pressures ranging from 100 to 1000 bar at constant flow and over a wide concentration range for three low-molecular-weight solutes, antipyrine, sodium 2-naphthalenesulfonate, and benzyltriethylammonium chloride, on an Eternity C18 stationary phase. The adsorption energy distribution was bimodal for all solutes, remaining clearly so at all pressures. The bi-Langmuir model best described the adsorption in these systems and two types of adsorption sites were identified, one with a low and another with a high energy of interaction. Evidence exists that the low-energy interactions occur at the interface between the mobile and stationary phases and that the high-energy interactions occur nearer the silica surface, deeper in the C18 layer. The contribution of each type of adsorption site to the retention factor was calculated and the change in solute molar volume from the mobile to stationary phase during the adsorption process was estimated for each type of site. The change in solute molar volume was 2-4 times larger at the high-energy site, likely because of the greater loss of solute solvation layer when penetrating deeper into the C18 layer. The association equilibrium constant increased with increasing pressure while the saturation capacity of the low-energy site remained almost unchanged. The observed increase in saturation capacity for the high-energy site did not affect the column loading capacity, which was almost identical at 50- and 950-bar pressure drops over the column. (C) 2016 Elsevier B.V. All rights reserved.