Predicting drug solubility in organic solvents mixtures: A machine-learning approach supported by high-throughput experimentation.

A novel approach based on supervised machine-learning is proposed to predict the solubility of drugs and drug-like molecules in mixtures of organic solvents. Similar to quantitative structure-property relationship (QSPR) models, different solvent types are identified by molecular descriptors, which, in this study, are considered as UNIFAC subgroups. To overcome the potential lack of UNIFAC subgroups for the complex Active Pharmaceutical Ingredients (APIs) currently developed in the pharmaceutical industry, the API molecule is considered as a unique entity in the proposed modelling approach. Therefore, API solubility is predicted as a function of temperature, functional subgroups of the solvents and composition of the solvent mixture; in turn, regressors' correlation is handled through Partial Least-Squares (PLS) regression. The method is developed and tested with experimental data of a real API and 14 organic solvents that are industrially employed for crystallisation. Solubility predictions are accurate and precise for single solvents, binary mixtures and ternary mixtures of organic solvents at different compositions and temperatures, with a determination coefficient R2 ≥ 0.90. To further test the applicability of the model, the proposed approach is applied to 9 literature organic solubility datasets of drugs and drug-like compounds and compared to benchmark solubility models in the literature. Results show that the proposed approach provides satisfactory predictions: the majority of validation and calibration data have R2 = 0.95-0.99; the ratio between RMSE (root mean squared error) of the proposed method and the range of measured solubility values is from 1 to 3 orders of magnitude smaller than the RMSE ratio obtained by the benchmark models.

Investigation of the Formaldehyde-Catalyzed NNitrosation of Dialkyl Amines: An Automated Experimental and Kinetic Modelling Study Using Dibutylamine.

The potential for drug substances and drug products to contain low levels of N-nitrosamines is of continued interest to the pharmaceutical industry and regulatory authorities. Acid-promoted nitrosation mechanisms in solution have been investigated widely in the literature and are supported by kinetic modelling studies. Carbonyl compounds, particularly formaldehyde, which may be present as impurities in excipients and drug product packaging components or introduced during drug substance manufacturing processes are also known to catalyze nitrosation, but their impact on the risk of N-nitrosamine formation has not been systematically investigated to date. In this study, we experimentally investigated the multivariate impact of formaldehyde, nitrite and pH on N-nitrosation in aqueous solution using dibutylamine as a model amine. We augmented a published kinetic model by adding formaldehyde-catalyzed nitrosation reactions. We validated the new kinetic model vs. the experimental data and then used the model to systematically investigate the impact of formaldehyde levels on N-nitrosamine formation. Simulations of aqueous solution systems show that at low formaldehyde levels the formaldehyde-catalyzed mechanisms are insignificant in comparison to other routes. However, formaldehyde-catalyzed mechanisms can become more significant at neutral and high pH under higher formaldehyde levels. Model-based sensitivity analysis demonstrated that under high nitrite levels and low formaldehyde levels (where the rate of formaldehyde-catalyzed nitrosation is low compared to the acid-promoted pathways) the model can be used with kinetic parameters for model amines in the literature without performing additional experiments to fit amine-specific parameters. For other combinations of reaction parameters containing formaldehyde, the formaldehyde-catalyzed kinetics are non-negligible, and thus it is advised that, under such conditions, additional experiments should be conducted to reliably use the model.

Kinetic Barriers to Disproportionation of Salts of Weakly Basic Drugs.

Disproportionation is a major issue in formulations containing salts of weakly basic drugs. Despite considerable interest in risk assessment approaches for disproportionation, the prediction of salt-to-base conversion remains challenging. Recent studies have highlighted several confounding factors other than pHmax that appear to play an important role in salt disproportionation and have suggested that kinetic barriers need to be considered in addition to the thermodynamic driving force when assessing the risk of a salt to undergo conversion to parent free base. Herein, we describe the concurrent application of in situ Raman spectroscopy and pH monitoring to investigate the disproportionation kinetics of three model salts, pioglitazone hydrochloride, sorafenib tosylate, and atazanavir sulfate, in aqueous slurries. We found that even for favorable thermodynamic conditions (i.e., pH ≫ pHmax), disproportionation kinetics of the salts were very different despite each system having a similar pHmax. The importance of free base nucleation kinetics was highlighted by the observation that the disproportionation conversion time in the slurries showed the same trend as the free base nucleation induction time. Pioglitazone hydrochloride, with a free base induction time of <1 min, rapidly converted to the free base in slurry experiments. In contrast, atazanavir sulfate, where the free base induction time was much longer, took several hours to undergo disproportionation in the slurry for pH ≫ pHmax. Additionally, we altered an established thermodynamically based modeling framework to account for kinetic effects (representing the nucleation kinetic barrier) to estimate the solid-state stability of salt formulations. In conclusion, a solution-based thermodynamic model is mechanistically appropriate to predict salt disproportionation in a solid-state formulation, when kinetic barriers are also taken into consideration.

Design Space Identification and Visualization for Continuous Pharmaceutical Manufacturing.

Progress in continuous flow chemistry over the past two decades has facilitated significant developments in the flow synthesis of a wide variety of Active Pharmaceutical Ingredients (APIs), the foundation of Continuous Pharmaceutical Manufacturing (CPM), which has gained interest for its potential to reduce material usage, energy and costs and the ability to access novel processing windows that would be otherwise hazardous if operated via traditional batch techniques. Design space investigation of manufacturing processes is a useful task in elucidating attainable regions of process performance and product quality attributes that can allow insight into process design and optimization prior to costly experimental campaigns and pilot plant studies. This study discusses recent demonstrations from the literature on design space investigation and visualization for continuous API production and highlights attainable regions of recoveries, material efficiencies, flowsheet complexity and cost components for upstream (reaction + separation) via modeling, simulation and nonlinear optimization, providing insight into optimal CPM operation.

Acute right atrial and pulmonary artery bone cement mass emboli following vertebroplasty.

Cardiac and pulmonary artery emboli are lethal complications following vertebroplasty. Clinicians should recognise these fatal complications immediately and surgical extraction is mandatory and provides the best outcome.