Saujanya Gosangari, Ph.D, director, technical services, Alcami rethinks risk in OSD process transfers.
Alcami
Technology transfer in oral solid dose (OSD) manufacturing is often framed as a structured comparison of materials, equipment, process parameters, analytical methods, and acceptance criteria. Yet even when these elements appear aligned, subtle differences in equipment behaviour or formulation sensitivity can reveal themselves only at commercialscale qualification.
A recent case study involving the transfer of a common blend process from a Vblender to a double cone blender illustrates why process understanding must extend beyond engineering batch success and why blend sampling remains an essential diagnostic tool.
A reasonable transfer — But not an assumption
The product used a common blend approach to support multiple dosage strengths. This is an efficient strategy, but it also places increased importance on blend uniformity because the same blend supports more than one strength. Any meaningful distribution issue within the blend could therefore have broader impact across the product presentation.
At the sending site, the product was manufactured using a V-blender. At the receiving site, the available equipment was a double cone blender. Both are well-established tumble blending technologies, and the transfer appeared reasonable from a high-level process perspective. The receiving site had appropriate capacity, the unit operation remained dry powder blending, and the overall manufacturing process was not fundamentally changed.
However, the transfer strategy did not assume that the two blenders would behave identically. While both pieces of equipment rely on tumbling motion, their internal geometry and powder movement patterns differ. The receiving-site team recognised that fill level, blend time, rotation speed, order of addition, material characteristics, and blender design could influence blend behaviour. For this reason, the transfer approach included engineering evaluation, targeted sampling, finished product assessment, and additional technology transfer batches to characterise performance before final process performance qualification (PPQ).
Initial engineering batches provided a useful level of confidence. The process was manufacturable, the blend could be discharged and processed and finished product results were acceptable. These data supported continued progression of the transfer, but they were not treated as the end of process understanding. Instead, they served as the foundation for further evaluation.
Technology transfer batches as a diagnostic tool
As part of the transfer strategy, additional technology transfer batches were executed to better understand how the formulation behaved in the receiving-site equipment train. These batches were not viewed merely as confirmation runs. They were used to challenge assumptions, evaluate equipment-related behaviour, and generate data that could inform the final qualification strategy.
From a manufacturing standpoint, the process performed acceptably. The blend was produced, discharged, and used to manufacture the applicable dosage strengths. Stratified content uniformity results for finished tablets were acceptable and did not indicate dosage unit uniformity concerns.
Blend sampling, however, provided additional insight. Samples collected from different locations within the double cone blender showed a location-dependent assay trend. The results suggested that the formulation exhibited some degree of spatial distribution within the receiving-site blender. Importantly, this observation was identified during the technology transfer phase, prior to final PPQ execution.
Because the observation was detected before final qualification, the team had the opportunity to evaluate the data, understand the likely cause, and determine whether the existing control strategy remained appropriate.
The importance of blender geometry
One of the most instructive aspects of this case is the reminder that not all tumble blenders behave the same way. While both Vblenders and double cone blenders rely on tumbling motion, the mechanisms by which they move powder differ in meaningful ways. A Vblender promotes repeated splitting and recombining of the powder bed as material flows through the two arms of the vessel. This motion can be particularly effective for formulations with low dose components or materials sensitive to distribution.
A double cone blender, by contrast, relies more heavily on cascading, rolling, and tumbling across its conical geometry. Depending on fill level, rotation speed, blend time, and material properties, certain regions of the powder bed may experience different movement patterns. For many products, these differences may not matter. For a low-dose, blend-sensitive, or segregation-prone formulation, they can become important.
In this case, the additional technology transfer batches helped reveal that the receiving-site blender produced acceptable finished product but also showed a measurable location-dependent trend at the blend sampling stage. This underscores a key lesson: even when equipment appears equivalent in principle, formulationspecific behaviour may diverge.
Investigating the source of the bias
The investigation focused on determining whether the PPQ blend results reflected analytical variation, sampling error, incomplete mixing, segregation, or equipmentrelated spatial distribution. The team reviewed blend parameters, fill volume, blend time, rotation speed, order of addition, material characteristics, sampling technique, and discharge process.
A critical step was mapping results by blender location. When viewed individually, the data could be interpreted as normal variability. When grouped by location, however, a clear pattern emerged: certain areas of the blender were consistently different from others, thus supporting the conclusion that the observation was related to spatial distribution rather than analytical scatter.
This type of locationbased mapping is an underused but powerful tool in OSD process evaluation. It can reveal subtle distribution patterns that would otherwise remain hidden.
Why SCU still passed
The fact that SCU remained acceptable across multiple strengths is not contradictory. Blend sampling and SCU measure different aspects of the process. Blend sampling captures the state of the powder at a specific moment and location. SCU reflects the cumulative effect of downstream handling.
Downstream steps such as hopper transfer and feed frame movement can redistribute material, reducing the impact of localised blend differences. This does not negate the importance of blend sampling; rather, it highlights the need to interpret both tests together.
Building robust OSD transfer strategies
This case reinforces the importance of process understanding in OSD technology transfer. A successful transfer requires more than matching equipment capacity or replicating blend times. It requires understanding how a specific formulation behaves within a specific equipment train.
For blend‑sensitive products, this includes:
- evaluating blender geometry and movement patterns
- designing sampling plans that can detect spatial trends
- interpreting blend and finished product data together
- considering how downstream steps may redistribute material
- ensuring that engineering batches are representative but not definitive
Conclusion
The transfer from a V‑blender to a double cone blender was technically reasonable, and engineering batches showed no evidence of blend bias. Finished product SCU results were acceptable across multiple strengths. Yet blend sampling in subsequent batches revealed a location‑based trend that required further evaluation. This case illustrates why OSD technology transfer must incorporate equipment‑specific understanding, robust sampling strategies, and a willingness to investigate unexpected findings.
Importantly, the issue was identified and addressed before final PPQ. The final PPQ was successful, with acceptable finished product performance across the applicable strengths.
Blend behaviour is formulation‑specific, equipment‑specific, and sometimes only fully revealed with additional technology transfer batches. Recognising this reality, and designing transfer strategies, accordingly, remains essential to building robust, reproducible commercial processes.