Novel aspects for the analysis and optimisation of amorphous formulations

Abstract

Formulating drugs in the amorphous form is an attractive and promising strategy to overcome the poorly water soluble challenge commonly encountered in the newly discovered drug. The active research in the investigation of the properties and performance of drugs in the amorphous form has revealed the major challenges in developing amorphous formulations into marketable products. The three main challenges are development of analytical techniques to characterise the amorphous formulations, prediction of performance and the design of amorphous formulations. The main aim of this thesis was to generate knowledge to overcome some of the challenges in the development of amorphous formulations with optimal performance. The first part of this thesis explored the use of two novel analytical techniques (i.e. combination of partial least square (PLS) analysis and low frequency Raman spectroscopy and sum frequency generation (SFG) imaging) in detecting low levels of crystallinity. Then, the ability of various conventional analytical techniques in predicting the critical quality attributes (CQAs) of amorphous formulations was investigated. Finally, the effect of commonly used excipients on the dissolution profiles of amorphous formulations was studied.

Low frequency Raman spectroscopy was successfully used in combination with PLS to quantify low levels of crystallinity in amorphous griseofulvin tablets during storage. The root mean square error of prediction (RMSEP) values of the PLS model based on the low frequency Raman spectroscopy was low (i.e. 1.14%) and was not very different from the RMSEP values of the PLS models built on the mid-frequency Raman regions collected using two different types of instruments (i.e. one which was the same instrumental setup as for the low frequency Raman regions and the other one was a FT-Raman spectrometer). The PLS models of the three types of Raman techniques were used to construct the recrystallisation profiles of neat amorphous griseofulvin tablets during storage. It was found that the recrsytallisation profiles based on the three different types of Raman techniques showed a similar trend. However, they differed in detecting onsets of crystallisation. Amongst the three techniques studied, low frequency Raman spectroscopy detected the earliest onset of crystallisation. The combination of PLS and low frequency Raman spectroscopy offers a new approach in quantifying active pharmaceutical ingredients (APIs) and different solid state forms in pharmaceutical systems. SFG imaging was successfully used to image the distribution of crystallinity in tablets containing neat amorphous griseofulvin. SFG imaging demonstrated that crystallisation on the surface of the tablet occurred upon compression whilst only minimal crystallisation was found in the bulk of the tablets. Storing the tablets resulted in the progression of crystallisation predominantly on the surface of the tablets. It was found that SFG imaging was able to detect crystallinity on the surfaces of the tablets upon compression when ATR-FTIR spectroscopy and SEM imaging were not able to do that. The ability of SFG imaging to detect crystallinity in amorphous griseofulvin tablets containing an excipient was also demonstrated in this thesis. Excipients that were used included nanoparticulate silica, HPMC, MCC and PEG and they were all SFG inactive. It was found that the presence of some excipients modified the crystallisation rate of amorphous griseofulvin. Thus, it is important to select the appropriate excipients for the development of amorphous formulations with optimal physical stability. In conclusion, SFG imaging is an attractive technique for imaging the distribution of crystallinity in an amorphous matrix due to its many advantageous features over conventional imaging techniques.

The ability of X-ray powder diffractometry (XRPD), Raman spectroscopy coupled with principal component analysis (PCA) and differential scanning calorimetry (DSC) in predicting the CQAs (i.e. physical stability and dissolution behaviour) of milled amorphous glibenclamide samples was investigated. Glibenclamide was milled for various durations and analysed using the three different techniques. It was found that the different analytical techniques suggested glibenclamide was fully amorphous after different milling durations. Glibenclamide was X-ray amorphous after 30 min of milling and was indicated to be fully amorphous by Raman spectroscopy after 60 min of milling. The onset of crystallisation obtained using DSC indicated glibenclamide was fully amorphous after the longest milling duration (i.e. 120 or 150 min, depending on the source of the glibenclamide used). It was found that the onset of crystallisation obtained using DSC best predicted the physical stability and dissolution behaviour (in sink and non-sink conditions) of milled amorphous glibenclamide samples. The postulated reason for the observations is due to the residual nuclei that were present in the samples milled for longer milling durations being able to result in a change in the onset of crystallisation (obtained using DSC), but had no effect on the XRPD and Raman signals. These residual nuclei (not detected by XRPD and Raman spectroscopy) acted as seeds for recrystallisation during storage and dissolution. In conclusion, the CQAs should be always considered when selecting analytical techniques to characterise amorphous formulations during the development and manufacturing stages of amorphous formulations.

The effect of various excipients and the method of introducing the excipients on the dissolution performance of amorphous indomethacin formulations in non-sink conditions were investigated. The excipients investigated were polymers (i.e. HPMC and PVP) and surfactants (i.e. Tween 80 and poloxamer 407) and binary combinations of the polymer and surfactant. The presence of surfactants predissolved in the dissolution medium was detrimental to the dissolution of neat amorphous indomethacin, suggesting that the surfactants promoted the crystallisation of amorphous indomethacin. Predissolved polymers if used alone were effective in maintaining supersaturation during dissolution of neat amorphous indomethacin. However, the use of predissolved binary combinations of polymer and surfactant were not able to maintain supersaturation during dissolution of neat amorphous indomethacin. The results were different when the polymer was incorporated in the solid amorphous formulations (i.e. solid dispersions) and the surfactants were predissolved in the dissolution medium. The dissolution of selected amorphous polymer based solid dispersions in the presence of predissolved surfactants resulted in an initial high degree of supersaturation which decreased gradually overtime. Even though supersaturation was not maintained throughout the experimental period (8 hours), the transient dissolution benefits of the surfactants on polymer based solid dispersions can be exploited in drugs which are predominantly absorbed in the initial part of the gastrointestinal tract. This study also demonstrated that the types of polymer and surfactant have an effect on the dissolution behaviour of the amorphous formulation. Overall, this study showed that the design of amorphous indomethacin formulations with optimal dissolution performance requires the appropriate selection of a combination of excipients and method of introducing the excipients.

It can be concluded from this thesis that the successful development of marketable amorphous formulations is not an easy task. It requires the rational consideration of many factors which include the selection of appropriate analytical techniques for the evaluation of amorphous formulations, development of methods to predict the CQAs of the amorphous formulations and the rational design of amorphous formulations with optimal performance.