Feed frame optimisation for multi-unit pellet system blends in a rotary tablet press

Article by Federica Giatti, Lorenzo Menarini, Caterina Funaro, Fabrizio Consoli, and Fabiano Ferrini at IMA Active.

Oral drug administration is the most convenient administration route for patients. Oral products include single-unit dosage forms (SUDFs) and multiple-unit dosage forms (MUDFs). SUDFs have a single core, which may contain the active pharmaceutical ingredient (API) mixed with excipients or may be neutral and coated with a layer containing the API. MUDFs consist of small particles that are typically contained in a capsule or a tablet. Their most important benefit is that the small particles more evenly distribute in the gastrointestinal tract than a single core does, resulting in less variance in transit time through the gastrointestinal tract compared to SUDFs.

MUDFs offer more predictable gastric emptying, less dependence on the patient’s state of nutrition, a high degree of dispersion in the digestive tract, less absorption variability, and a lower risk of dose dumping. As a consequence, MUDFs generally avoid high local drug concentration and local irritation. In addition, MUDFs can improve patient compliance because of ease of swallowing and flexible dose fractionation delivering incompatible drugs or excipients thanks to differences in coating application. 

Moreover, MUDFs represent an innovative way to introduce drugs with different release profiles answering the need of a multi-target approach in treating different diseases. Last but not least, converting a single unit dosage form in MUDFs extend the commercial value of a given drug.

The multi-unit pellet system (MUPS) is one of the most popular MUDFs and is generally filled into a capsule or compressed with other excipients to form a tablet. MUPS tablets (TMUPS), in particular, are composed of coated or uncoated pellets with a powder phase and excipients as necessary for compaction into a tablet. 

The API may be included in the core of the pellets (polymeric matrix) or layered onto the pellets with suitable excipients for modified-release purposes (delayed or sustained). Delayed release is generally achieved using polymers such as methacrylic acid copolymers or hypromellose (HPMC). Sustained release is obtained by using dedicated polymers (such as ethyl cellulose) with different film thicknesses or by using uncoated pellets that act as a matrix polymeric system (such as a hydrophilic cellulosic matrix). The powder phase generally contains fillers, binders, disintegrant, and lubricant.

While innovative, TMUPS can present manufacturing challenges, with potential polymer damage and blend segregation being the most well-known issues. For this reason, only a few pharmaceutical TMUPS are commercially available on the market, as shown in Table 1.

Ideally, TMUPS should disintegrate rapidly in the gastrointestinal tract, guaranteeing the same drug release as an uncompressed MUPS powder. However, compression-induced damage to functional polymers could lead to undesired drug-release with adverse effects, and damage to esthetic polymers could alter their taste-masking function. Also, polymeric coatings must be protected from cracking when applying certain compaction pressures. 

Segregation can also be a concern. For a TMUPS formulation, the MUPS pellets should generally have a particle size distribution in the range of 300 to 2,000 micrometers and should make up between 20-70% of the formulation, whereas the other excipients will have an array of particle sizes smaller than 200 micrometers. This broad range of particle sizes can potentially result in size segregation during handling of the blend prior to and during tableting. If segregation occurs, the resulting tablets could fail to meet API homogeneity requirements and be rejected during quality control. Effective tablet press feeding system operation is critical to ensure adequate blend homogeneity and avoid out-of-specification tablets.

The following case study describes how to optimise the tablet press feed frame for a MUPS formulation to verify process feasibility on a rotary tablet press.

Materials and methods 

A MUPS placebo formulation based on a customer’s existing formulation on the market was used as a reference. The formulation was composed of pellets (Suglets, Colorcon) and other excipients necessary for compression, as shown in Table 2. The pellets had an average particle size of 400 micrometers and were coated with an ethyl cellulose-based polymer (Surelease, Colorcon) with a blue colorant (FD&C, Colorcon) added to simplify their identification inside the tablets. Pellet coating was performed in bottom-spray configuration fluid-bed equipment (Aria 120, IMA). The final coated pellets were mixed with microcrystalline cellulose (MCC) and magnesium stearate in a tumble blender with a 25-liter bin (Cyclops, IMA), with the mixing endpoint determined through NIR spectroscopy. The NIR instrument was a MicroNIR PAT-W (Viavi) equipped with a 128-pixel InGaAs photodiode array with a wavelength range of 950 to 1,650 nanometers (10,526 to 6,060 cm-1).

The MUPS formulation was then loaded into a rotary tablet press (Prexima 300, IMA). The bin of the tumble blender was fastened directly onto the tablet press hopper to ensure as little vibration as possible and minimise segregation of the MUPS blend. 

The tablet press was equipped with a 27-station Euro-D turret with 13-millimetre round, flat-faced punches. The punches were keyed and had logos on both on the upper and lower punch to identify their orientation with respect to turret movement, as shown in Figure 1. 

A response surface method (RSM) design of experiments approach was followed to deeply investigate the influence of critical process parameters and evaluate tableting feasibility. The critical variables selected were feeding system type (three-paddle die feeder [3PDF] two-paddle die feeder [2PDF], high die feeder [HDF], standard die feeder [SDF], and gravity die feeder [GDF]); loading paddle type (flat-slanted, increased-volume flat, flat, round, round off-set, and V-shaped); single paddle speed (15 to 135 rpm); and turret speed (20 to 80 rpm), as summarised in Table 3.

Analytically, a specific method was developed to count the percentage of pellets inside each tablet and their concentration deviation. Through Design-Expert software, the results were analysed by using the analysis-of-variance (ANOVA) method to determine the significant portion of variance. After checking that the model was significant, the different geometric variables (first feeder type, then paddle profile) were compared, focusing on the interaction for each design point for the same process variables (turret and paddle speed). 

To evaluate process reliability and enforce results obtained with the analytical method, samples were also analysed visually. The focus was on pellet distribution on each tablet face—the presence of logos, as shown in Figure 2, was used to characterise whether a specific face or tablet quarter included undesired pellet agglomeration or whether the pellets were correctly distributed.

The target tablet weight was 500 milligrams, with an 8-kilonewton main compression force (optimised to avoid breaking the pellet coating). The precompression chamber, upper punch penetration, and loading cam were kept constant.

Results and discussion

The final mixture was obtained thanks to smooth and reproducible granulation and mixing step: blending endpoint was evaluated through MicroNir Pro 3.0 software by Viavi. The raw spectra were pretreated using the Savitzky-Golay first derivative (5 points) and then the standard normal variate on the entire spectral range. Using moving block standard deviation (block size = 30), an example of the behaviour of the mixing stage is shown in Figure 3. As the figure shows, homogeneity (under the acceptable level of standard deviation) was reached both immediately before adding the lubricant and at the end of the overall mixing stage. Tableting was possible with all the mechanical configurations applied, and weight stability was reached without any issues.

A mixture of analytical and visual results were obtained. The discrepancy in terms of pellet concentration expected (43%) versus actual and the relative standard deviation in pellet concentration when analysing tablets within the same test, are shown in Figures 3 and 4.

Figure 3 represents the variance for all the feeder types mounted on the machine with different loading paddle configurations versus pellet concentration. HDF and 3PDF provided values closest to the formulation’s pellet concentration (43%) with low standard deviations on pellet concentration.

Figure 4 represents one factor plot for pellet

concentration variation for different paddle profiles in the feeder types analysed. The flat profile provided the lowest standard deviation, close to the 43% standard.

Visual inspection was the key factor to enforce previous results. The HDF feeder (Figure 5) achieved the best results, as the die filling is less dependent on mechanical parts, which can induce segregation. The SDF and 2DPF feeders also achieved acceptable results, as shown in Table 4.

The same approach was taken for paddle shape, which has a clear influence on the process. The flat paddle profile (Figure 6 right) achieved the optimal balance between analytical and visual results and avoided segregation better than the round profile (Figure 6 left), as shown in Table 5. The flat slanted paddle profile also produced tablets with an acceptable pellet distribution. 

Paddle speed has an important influence on the results obtained. The position inside the die feeder influences the functionality of each paddle. The paddle on the left, called the loading paddle, is more responsible for loading the dies, while the paddle on the right, called the dosing paddle, tends to complete the die filling and scrape off the excess powder.

Setting the loading paddle speed to almost twice the turret speed maintained the homogeneity of the MUPS blend over time. 

The dosing paddle speed is related to the back-dosing of the blend inside the feeder. When passing below the feeding system, the lower punch is initially completely down to fill the die. Then, the punch moves up to fine tune the dosage to obtain the required tablet weight. At this precise moment, the dosing paddle scrapes away the excess powder, back-mixing it inside the feeder. For this reason, considering the presence of particles with different densities and sizes, choosing a low speed prevents undesired scraping that can cause segregation. 

Predictably, high output in terms of tablet press speed negatively affects segregation, as centrifugal force-phenomena could begin to occur and cause segregation inside the dies.

Conclusions

MUPS blends are generating high interest from both pharmaceutical companies and equipment suppliers, though only a small number of TMUPS products are currently on the market. The main reason for this is the high number of challenges, both from a formulation perspective (such as pellet core, polymeric coating, and cushioning excipients) and a manufacturability perspective (such as feeding, balancing compression force and pellet integrity, and segregation). 

Avoiding segregation is mandatory for TMUPS manufacturing, but correct handling of the blend during pretableting processes along with correct feeder design has demonstrated the feasibility of obtaining acceptable TMUPS products.

A high-die feeder with flat-profiled paddles provided the best results, because this configuration moves the powder gently without negatively affecting blend homogeneity. 

Satisfying TMUPS manufacturing requirements also requires low compression forces to guarantee the integrity of functional pellet coatings, and the loading cam should be near in terms of millimeters to the dosing chamber chosen.