(W1130-09-61) Low Dose Formulations in Continuous Manufacturing of Tablets via Semi-continuous Feeding and Blending

Purpose: Despite years of efforts by industry and academia alike, extremely low dose formulations have been a troublesome part of the transition from batch to continuous manufacturing (CM) of oral solid dosages. A common starting point for CM is to operate loss in weight (LIW) feeders and an inline blender in a continuous fashion. Achieving the correct mass flow rates for each component is key to dispensing the correct unit formula and thus a successful operation. This makes controlling the individual material feeders crucial to maintaining a state of control. Many OSD CM applications operate at a throughput of approximately 25 kg/hr., which for an API at 1% concentration, correlates to just 4.2 g/min feeding. This, paired with the poor feedability of many APIs, makes for a difficult process control situation. One response to this and other barriers to the conversion to CM has been the introduction of alternative dosing techniques using a semi-continuous approach. The key difference with semi-continuous dosing is that the feeders charge a target weight per cycle, rather than a target mass flow rate. The following study's purpose was to show the feasibility of low dose formulations in CM with a semi-continuous feeding and blending approach.

Methods: A semi-continuous feeding-blending unit (Figure 1) was used to test a low dose formulation (Table 1) of micronized acetaminophen. The experimental setup consists of LIW feeders operating in a batch dispensing mode and a conical ribbon blender. Each feeder was loaded with a single formulation component and operates in a fully automated, semi-continuous manner with the downstream blender and blends were charged in to one of five parallel fluid bed granulators. Each sub-batch produced by the blender and subsequently granulated in a single fluid bed is called an “x-key”. After granulation and prior to the rotary tablet press, a secondary duplicate feeder-blender unit was used to feed and blend the lubricant with each x-key of granules.

Results: The process ran smoothly for four days straight of operation. In total, 720 x-keys were produced, corresponding to 2.16 tons of final tablets. The API was dosed with a 99.97% accuracy and standard deviation (SD) of 1.03% (Table 1). Much like fully continuous feeding in other systems, fed accuracy over time is the best indicator of resulting product content uniformity. The average UDU AV was 4.37% with a SD of 1.47% (Figure 2). This is far below the 15% upper specification limit and more than a 6σ tolerance, indicating a robust state of control and acceptable UDU throughout the continuous production of tablets. Other indicators of process health including product temperature at the end of spraying, granulation process drying time, x-key yield, and tablet physical characteristics all showed a consistent state of control as well.

Conclusion: A 1% drug load formulation was tested using a semi-continuous feeding, blending, and granulating system for the continuous production of tablets. The trial demonstrated an accurate and precise control of the dispensed quantities and easily passed acceptance value requirements for the uniformity of dosage units.

Acknowledgements: Abstract submitted by an equipment vendor, but nonetheless shows novel research in the space of low dose CM tableting.

Figure 1: Full CM line (left), feeder-blender unit (middle), dosing into blender (right)

Table 1: Formulation, process data, and trial results

Figure 2: Uniformity of Dosage Units