(W0930-08-55) Distributed Manufacturing of 3D Printed Pharmaceuticals: A Case Study

Purpose: Three-dimensional (3D) printing offers the potential to bring personalised medicine closer to the patient by enabling drug manufacturing at the point-of-care. Regulatory bodies, such as the US Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the UK Medicine and Healthcare products Regulatory Agency (MHRA), are now committed to address regulatory issues unique to point-of-care manufacturing. The MHRA has notably proposed various alternative decentralised models of manufacturing, including distributed manufacturing, which involves shifting the location of final dosage form assembly from centralised manufacturing sites to multiple delocalised locations. This study aims to examine the feasibility of a plausible supply chain and distributed manufacturing model for 3DP drug products. Losan Pharma GmbH (Germany) acted as the model pharmaceutical supplier, preparing feedstock material in the form of efavirenz granulates and delivering it to a distributed manufacturing site at the point-of-care (UCL School of Pharmacy). The granulates were directly used to fabricate efavirenz printlets using a direct powder extrusion (DPE) 3D printer. Subsequently, an in-line NIR system was developed to perform in-process quality control of the printlets. Physicochemical characterization and in vitro drug release studies were also performed to demonstrate the viability of the proposed model of manufacture.

Methods: Preparation of granulated formulations Efavirenz granulates were prepared at Losan Pharma GmbH under GMP protocols conditions (Table 1), kept in labelled plastic bottles and shipped to the UCL School of Pharmacy for further processing into DPE 3D printed tablets.
3D printing process The granulated formulations were printed using a DPE 3D printer (M3DIMAKER 1, FabRx, UK) (Table 1). The templates of the printlets were exported as a .stl file into 3D printer software and the selected 3D geometry was a round printlet (10 mm diameter × 3.6 mm height).
In-line NIR system The NIR system was a portable MicroNIR 1700ES near infrared spectrometer. To perform in-line data acquisition, the portable MicroNIR with a tablet probe was attached to the moving part of the 3D printer, parallel to the printer printhead and perpendicular to the building plate (Figure 1). A specialised automatically performed in-line data acquisition and collected reflectance spectra in triplicate. When the data acquisition is completed, the printhead moved to the next position to print the following printlet, and the process was repeated until the entire batch is printed.
Model development and multivariate data analysis Printlets prepared with eight concentrations of efavirenz (n=4) were used for model development (0%, 5%, 10%, 15%, 20%, 25%, 30%, 35%). Each printlet was analysed three times. Five printlet concentrations (n=4) were selected for training the model (0%, 5%, 15%, 25%, 35%), and three concentrations (n=4) were used for testing the model (10%, 20%, 30%). Partial least squares (PLS) regression was performed on the datasets to build calibration models. Following NIR analysis, each individual printlet was quantitatively analysed for drug content via HPLC.
Characterisation of the printlets The printlets were characterized by DSC and XRPD. Drug content was quantified by HPLC, and in vitro dissolution test was performed to obtain the drug release profiles. In vitro drug release profiles were obtained using a USP-II apparatus. The printlets (n = 4) were placed in water with 2% SDS as indicated in the USP monograph for efavirenz tablets under constant paddle stirring (100 rpm) at 37 °C.

Results: All the formulations were found suitable for DPE 3D printing. 6 PLS models developed with 3 different pre-treatment filters and their combinations. A 10-fold cross validation with 3 repeats was performed to determine the optimal number of latent variables (LVs) in the PLS model. The model with the higher linearity (R² test = 0.9833) and the lowest root mean square error (RMSE test = 1.0662%) was selected (Figure 2). The selected model used 2nd derivative (Savitzky–Golay with a filter width of 25 and a 2nd polynomial) pre-processing technique. XPRD analysis indicated that efavirenz was in the crystalline form before printing. However, efavirenz existed in the amorphous form within the printlets. DSC results corroborated XPRD results indicating the amorphous form of efavirenz within the printlets. Finally, dissolution profiles showed that more than 80% of efavirenz was released in the first 60 min for all concentrations, reaching 100% of drug release within 120 min.

Conclusion: This study has successfully demonstrated the feasibility of a model of point-of-care manufacturing wherein pharmaceutical feedstock is prepared by a pharmaceutical industry partner under GMP, delivered to and fabricated at the satellite manufacturing site, and quality controlled with an in-process PAT tool (NIR spectroscopy). The 3D printing process and the acquisition of the printlets’ NIR spectra were both controlled by a unified software, facilitating accelerated throughput and batch release. The developed in-line dose verification system enables QC analysis to be performed immediately after fabrication, facilitating high throughput and real-time batch release in clinics in the future.

Acknowledgements: I.S.-V. acknowledges Consellería de Cultura, Educación e Universidade for her Postdoctoral Fellowship (Xunta de Galicia, Spain; ED481B-2021-019). Project partially funded by SmartDose project from Eureka Healthy ageing (Innovate UK, UK, 83030).

Table 1. Formulations composition.

Figure 1. On top, pictures of the pharmaceutical 3D printer (left) and detail of the DPE printhead with the attached NIR probe (right). Below, images of the different 3D printed tablets with varying concentrations of the drug efavirenz. Scale shown in cm.

Figure 2. PLS model of NIR predicted vs. HPLC determined efavirenz content. Grey points are calibration (5 concentrations) and black are validation (3 concentrations).