• Application of near-infrared spectroscopy to monitor in-process and product quality attributes for granulation and immediate release tablet production

      Kona, Ravikanth; Hoag, Stephen W. (2012)
      The purpose of this study was to investigate the application of NIR spectroscopy to monitor critical in-process and excipient variables in the manufacturing of immediate release tablets. In this study, roller compaction and fluid bed granulation were used for the granulation of Ciprofloxacin hydrochloride and Fexofenadine hydrochloride, respectively. In roller compaction, prior knowledge was systematically incorporated into the risk assessment using failure mode and effect analysis (FMEA). The factors identified using FMEA were quantitatively assessed using a Placket-Burman screening design. Results indicated that roll pressure (RP) was the most critical roller compaction process factor affecting the granule size and flow. Binder grade, Klucel® EXF vs JF and tablet compression force (Pmax) were found to be critical to the release characteristics of Ciprofloxacin. This study demonstrated that the scientific rationale and quality risk management analysis were used to successfully and efficiently determine the Critical quality attributes. Four factors, RP, Pmax, binder source (Klucel® EXF vs Nisso®-L), and lubricant type (monohydrate vs dihydrate), were further investigated using high resolution experimental design. The results showed that binder and lubricant replacement was insignificant. In addition, NIR was used at various stages of the product development, and the partial least square (PLS) regression models developed have successfully predicted blend uniformity, particle size, crushing force, and disintegration time of the batches manufactured in the same location. However, the models yielded higher prediction errors for batches manufactured at a different location. In fluid bed granulation, a PLS model was developed to successfully predict the moisture levels in real time using in-line NIR probe specially designed for this application. This combined with humidity and temperature data from a novel PyroButton® in conjunction with multivariate data analysis techniques was used to develop multivariate statistical process control charts (MSPC) such as scores, distance to model (DModX), and Hotelling T2. The application of these charts for process monitoring and fault detection was further evaluated. In summary, these findings demonstrate the application of NIR in conjunction with multivariate chemometric models as a potential Process Analytical Technology (PAT) tool for effective process monitoring and process control in pharmaceutical manufacturing.
    • Development of Real Time Release Testing of Controlled Release Multiparticulate Drug Delivery System Using the Principles of Quality by Design

      Kothari, Bhaveshkumar H.; Hoag, Stephen W. (2014)
      To develop a control system, for real time release testing of a controlled release multiparticulate drug delivery system, an adequate understanding of polymer film formation and effect of processing parameters on the quality of film formation is essential. Presently, the curing of pseudolatex films is not well understood, and without a proper understanding of film formation mechanisms products without highly variable dissolution cannot be developed. To better understand film formation, the material attributes and process parameters were systematically assessed using risk analysis models like Ishikawa and failure mode and effect analysis (FMEA). This was followed up by a resolution V fractional factorial design to gain process understanding. Information gained was further evaluated using a resolution IV fractional factorial design to identify the critical process parameters that can significantly influence drug dissolution due to poor film formation. The design space was evaluated using different statistical approaches and experiments were conducted using central composite response surface methodology design to map the response surface and determine edge of failure. The in-process control strategy models were developed using diffuse reflectance near infrared spectroscopic technique. The risk assessment models and the statistical experimental designs helped to elucidate the effect of process efficiency and variation of extent of curing during the coating process. The design space was established using two different statistical models and were in close agreement to each other with statistical least square approach being more conservative than the Bayesian approach. The coating process was optimized and design space was built with product temperature, curing temperature and curing time deemed as the most critical process parameters. The effect of humidity on the extent of curing was also characterized and the in-process control strategy models helped determine process trajectory which could serve as the basis for a process control chart and actual endpoint measurement of the coating process. The intrinsic process variability associated with the coating process was successfully studied and in-process models were developed using near infrared spectroscopy and the data fusion method provided new insights into the prediction of dissolution from coated beads.