Fabiana Moscovici

For several years now, the pharmaceutical industry has been hearing more about Industry 4.0 (1), which allows massive amounts of data to be collected, aggregated, and analyzed in real time and then used for predictive and other types of modeling. Manufacturers in a broad range of industries have embraced Industry 4.0 concepts and recognized its potential to help them respond more nimbly to changes, predict future trends, fix problems without stopping production, and trace issues back to root causes (2). 

	Industry 4.0 depends on Big Data to produce results and support quick decision-making (3,4). If pharmaceutical quality by design (QbD) is the methodological basis for developing robust processes, Industry 4.0 is the paradigm shift required to introduce innovative technology in pharmaceutical manufacturing. Its goal is to support decision-making based on the comprehensive information extracted from a full exploration of available data.

	Manufacturers must go through evolutionary change in order to be ready to implement Industry 4.0. In considering readiness, senior managers should ask:

		What operational issues would be mitigated by taking an Industry 4.0 approach to manufacturing?

		How can the move to Industry 4.0 be made most efficiently?

		Does the company have the infrastructure required for implementing the tools and technologies needed (e.g., process analytical technologies [PAT] and more advanced process control)?

	There is no point in connecting inefficient processes that cannot guarantee the required level of product quality. Existing process bottlenecks and quality concerns should be addressed before attempting to enable existing systems for Industry 4.0. Processes should be built with the vision of generating knowledge along each step of the pharma value chain. Ideally, manufacturers should use QbD methods to develop feedback loops for passing information along, so that, at any point along the way, the process is ready for the next step (4). 

	This article describes how QbD principles and real-time process monitoring, via a PAT tool, were used to optimize a unit operation used in oral solid dosage (OSD) manufacturing. This work involved collaboration between 4Tune Engineering, a consulting firm with more than 16 years of experience working in pharma and biopharma, and Aché Pharmaceutical Laboratories, whose leaders are moving toward the complete adoption of Industry 4.0 principles in manufacturing.

	The process involved was made up of six unit operations. The final operation was a two-step coating process that, on its own, accounted for half of the total process time and led to productivity losses and extended working hours. An analysis of historical process data revealed which process parameters had the greatest impact on the duration of the two coating steps in that unit operation. 

	A subsequent criticality analysis identified which critical quality attributes (CQAs) were most affected by the coating and its process parameters. Performing this assessment resulted in an extensive knowledge base that allowed the team to describe the process and pinpoint where quality and productivity would overlap. Generating this information, in an organized form, was crucial to reducing process time without compromising product quality requirements. The use of small-scale models allowed larger ranges to be tested so that the optimal applicable values could be found without disrupting production schedules. 

	Through the application of design of experiments (DoE), relationships between process parameters, coating times, and CQAs were established and quantified. In all, 22 DoE experiments were carried out. Results pointed out the prospect of accentuated time reduction. By applying an optimization methodology, operating conditions could be tuned so that product quality was within specifications and coating time was reduced considerably. 

	To find the design space, the multivariate zone in which the simultaneous variation of parameters will guarantee product quality and short coating times, 5000 simulations were run with a normally distributed variation. The design space and optimized runs were tested at commercial scale, resulting in an approximately 25–51% reduction in the coating time required (

	Figure 1. Total coating time is reduced in the design space while maintaining an in-specification quality profile.

	Ensuring consistent product quality throughout the process is crucial, not only for process optimization but also for routine manufacturing. A real-time monitoring system allows any quality concerns to be verified or detected quickly.

	To ensure the required pharmacokinetics, the OSD form was coated with specific functional layers. These coats protect the API from being released in a low pH environment and guarantee the appropriated dissolution profile. In this way, the CQAs for enteric and dissolution profiles were evaluated by collecting samples and using conventional analytical methods (e.g., high-pressure liquid chromatography [HPLC]). 

	However, this approach is time and resource consuming, and the time gap between sampling and generating results  does not allow quality issues to be fixed immediately. To make this procedure more agile, a near-infrared (NIR) spectrometer probe was installed in-line and set to collect spectra every 30 seconds during the two-step coating process. 

	Using the spectra, a real-time monitoring program was developed to ensure that enteric and dissolution profiles were within specifications. This program uses a cascading system for decision-making: first, a multivariate model follows the enteric profile over time until the 2% API release threshold is assured (

); then, a combination of two multivariate models is used to monitor the dissolution profile evolution (

	The dissolution curve model was fitted to Higuchi and Logistic (5) and other equations, while the enteric profile was modeled directly from the data. If any quality issue arose during these phases, the process could be stopped immedidately, permitting troubleshooting and avoiding resource waste until the issue is addressed. 

	The real-time monitoring program allows for a continuous evaluation of the coating steps, ensuring that the desired layers of thickness (i.e., measurements that result in the enteric and dissolution profiles being within specifications) with no sampling, no laboratory equipment occupation, no reagents needed, and no time gap for the identification of issues in the two CQAs being monitored. This approach constitutes a clean, fast, and agile solution for process automation.

	The collaboration between 4Tune Engineering and Aché Pharmaceutical Laboratories showed how an Industry 4.0 approach can be used to revise, and considerably improve, existing commercial operations. This project is an example of pharma’s first steps toward Industry 4.0, with the collection and analysis of big data for optimization and decision-making. These approaches will undoubtedly continue to shape the manufacturing world in the future. Pharmaceutical companies should pursue Industry 4.0 and other new strategies to bring about innovation and world-class operations in terms of operational excellence and quality compliance.

	The authors would like to acknowledge and thank all the members of Aché’s quality, R&D, production, and regulatory affairs departments for their efforts and dedication to this project.

	1. B. Marr, “What is Industry 4.0? Here’s a Super Easy Explanation for Anyone,” 

, September 2, 2018.
	2. P. Guilfoyle, “Pharma 4.0: Industry 4.0 Applied to Pharmaceutical Manufacturing,”  

, October 23, 2018.
	3. J. Markarian,  “Industry 4.0 in Biopharmaceutical Manufacturing,”

, July 1, 2018.
	4. N.Calnan, M.J.Lipa, P.E.Kane, J.C.Menezes, 

A Lifecycle Approach to Knowledge Excellence in the Biopharmaceutical Industry

, CRC Press (2017).
	5. K. Ramteke, P.A. Dighe, et al., 


	Vol. 44, No. 7
	July 2020
	Pages: 34-36

	When citing this article please refer to M. Testas et al, "Using Industry 4.0 to Optimize Oral Solid Dosage Form Manufacturing," 

Articles

Real-time monitoring and the principles of quality by design were used to optimize an OSD coating process.