(T1530-02-13) De-Risking Tablet Formulation Development Using Compaction Simulation

Purpose: Development of a tablet formulation can be a long and costly process without the use of rapid and material sparing techniques that help mitigate risks. In this case study, compaction simulation was leveraged to define the critical process parameters (CPP) during roller compaction (RC) and tableting to ensure meeting the drug product (DP) critical quality attributes (CQAs) across lab and pilot scales.

Methods: A film-coated tablet formulation was developed as a DP candidate for bioequivalence clinical studies. The formulation consisted of 20 % API, 44.75 % mannitol, 28% microcrystalline cellulose, 3 % fumaric acid, 2% sodium croscarmellose, 0.5 % colloidal silicon dioxide and 1.75 % magnesium stearate. R&D trials (600-g batch): Pre-blending of raw materials was carried on a 3.5 L blender and screened using a Frewitt® conical mill (800-µm screen). Intra-granular (IG) lubrication was conducted for 12 min at 17 rpm. For the RC simulation step, a Compaction Simulator (Medelpharm Styl’One® Evolution) was equipped with flat 20x10 mm punches and set to mimick a Gerteis mini-pactor® at 3 RPM roll speed, 3 mm gap and to achieve a compact solid fraction (SF) between 70 -75%. Compacts were milled on a hand-mill (Gerteis® ,1 mm square screen). Extra-granular (EG) lubrication was performed in a 2 L blender for 12 min at 17 rpm. Evaluation of the formulation suitability for a RC process was performed on the StylOne® equipped with 11.28 mm round flat punches on a broad pressure range (50 to 350 MPa). Subsequently, the Styl’One® was set to simulate a Korsch XL-100 at 20 rpm throughput using oval tooling (17.48 x 8.74 mm) to manufacture tablets within the tensile strength (TS) specifications (3.0 – 3.5 MPa). Finally, core tablets were coated on a 0.8 L pan at 4 % weight gain. Pilot-scale manufacturing (7-kg batch): Pre-blending of raw materials was carried on a 50 L bin and screened on a Comil (Quadro U10®, 813-µm screen). IG lubrication was conducted for 10 min at 15 rpm. For RC, a Gerteis mini-pactor® was run with a roll force of 7.5 kN/cm, 3 rpm roll speed and 3 mm gap followed by milling (1 mm square screen). EG lubrication was performed for 10 min at 15 rpm on a 25 L bin . For tableting, a Korsch XL-100 was set at 20 rpm turret speed and compression force adjusted to reach the TS specifications. Finally, core tablets were coated on an IMA Perfima Lab® (10 L pan) at same weight gain.

Results: Formulation suitability for a dry granulation process was assessed at the lab-scale by comparing its tabletability before and after RC simulation. As shown in Fig 1, the formulation exhibited a limited work-hardening phenomenon (loss of bonding capacity) after RC simulation. Overall, final blend tabletability remains satisfactory as the core tablet target TS is achieved from 180 MPa approximately. At the lab-scale, compacts were manufactured at 11.5 kN compression force resulting in a mean SF of 0.72. Following the Styl’One® hybrid modeling, the corresponding SCF on a Gerteis mini-pactor® would be 7.6 kN/cm approximately. During transfer at pilot-scale, the RC was set at 7.5 kN/cm and the obtained ribbon SF was 0.72. Final blend properties were also consistent between both scales in terms of bulk density, Carr index and Hausner ratio (see Fig 2). Furthermore, particle size distribution was also similar, with some differences on the extremes: < 90 µm (18% lab vs 10% pilot) and 700 to 1000 µm (20% lab vs 28% pilot). This difference could be explained as the manual milling at lab scale that might lead to higher friction, increasing the amount of fines. Finally, to achieve a tablet TS of 3.2 MPa with the resulting granules at both scales: the pressure required on the Korsh XL-100 was slightly higher than the simulation on the Styl’One®: 230 MPa (31 kN) vs. 200 MPa (27 kN), respectively, as shown in Fig 3. This might be attributed to the lubrication time difference during scale-up. A longer lubrication time was used at the pilot scale. For obtaining the equivalent tablet TS, a higher compression pressure was needed and the disintegration time was slightly longer (2 min at lab-scale vs 4 min at pilot-scale). Furthermore, the slight difference observed in PSD, with more fines at lab-scale might also increase the tablet mechanical strength. From a performance perspective, the dissolution profiles for the coated tablets at both scales remained similar and resulted in more than 90% of the API dissolved 30 min.

Conclusion: Tablet formulation development was de-risked using compaction simulation as a material-sparing approach for RC and tableting steps. The target CPPs were defined at the lab- scale and similar CQAs were achieved across the lab and pilot scales for the drug product demonstrating successful scale-up.

Figure 1: Lab-scale: Pre- and Final blends tabletability.

Figure 2: Dry granulation: Lab- vs Pilot-scale.

Figure 3: Tabletability: Lab- and pilot-scale.